Home

News: IBM backs PHP. Un-named analyst makes a jab at Java

  1. IBM has announced that they are collaborating with Zend, of PHP fame. An un-named analyst got into the mix claiming that IBM is disillusioned with Java. What a bold comment. Give me a break.
    One industry executive who requested not to be named said that IBM's push into PHP and scripting reflects IBM's disillusionment with the Java standardization process and the industry's inability to make Java very easy to use.

    "IBM's been so fed up with Java that they've been looking for alternatives for years," the executive said. "They want people to build applications quickly that tap into IBM back-ends...and with Java, it just isn't happening."

    For his part, Smith said that Java and PHP can be used for different tasks and said that IBM remains committed to Java.

    Read more: IBM backs open-source Web software

    In related news:
    Separately on Friday, IBM announced plans to submit about 30 projects to SourceForge.net, a Web site that hosts thousands of open-source projects. Included in the IBM donations are the Jikes product.

    Threaded Messages (247)

  2. WTF[ Go to top ]

    Please do not post such "an un-named analyst opinions".
    This site is not to discuss people's opinion - specially not-so-technical (read judgemental) ones.

    Lets not lower down the quality of TSS.
    Long live TSS !

    - Nitesh
  3. WTF[ Go to top ]

    Please do not post such "an un-named analyst opinions".This site is not to discuss people's opinion - specially not-so-technical (read judgemental) ones.Lets not lower down the quality of TSS.Long live TSS !- Nitesh
    +1
  4. WTF[ Go to top ]

    Please do not post such "an un-named analyst opinions".This site is not to discuss people's opinion - specially not-so-technical (read judgemental) ones.Lets not lower down the quality of TSS.Long live TSS !- Nitesh
    +1

    +2 Quit posting crap like this. I feel like I'm on Slashdot
  5. WTF[ Go to top ]

    +1
  6. WTF[ Go to top ]

    Please do not post such "an un-named analyst opinions".This site is not to discuss people's opinion - specially not-so-technical (read judgemental) ones.Lets not lower down the quality of TSS.Long live TSS !- Nitesh
    +1
    +2 Quit posting crap like this. I feel like I'm on Slashdot

    +3

    Is TSS' pageviews dropping?
  7. WTF[ Go to top ]

    +4
    There is no use with this discussion.


    Thanks,
    Surya
  8. We definitely need the backing of IBM for PHP to push the nice scripting language to corporate Web development arena.PHP incorporate the nice features of Java, and Perl. It will become the language of choice on the UNIX and LINUX and even Windows platform for Web development. Because the Java development requires just too much configuration headache.
  9. We definitely need the backing of IBM for PHP to push the nice scripting language to corporate Web development arena.PHP incorporate the nice features of Java, and Perl. It will become the language of choice on the UNIX and LINUX and even Windows platform for Web development. Because the Java development requires just too much configuration headache.

    I'm sorry but I strongly disagree. I recently spent far too much time dealing with the installation of a major PHP application, which involved complicated configurations and compilations of apache and PHP (both being highly mutually version-dependent in this case), followed by installation of a range of PHP PEAR modules, some of which no longer existed, and many of which were still in beta. If I need to install another PHP application, I could well find that my current setup is incompatible. Comparing this with the trivial process of installing Tomcat and then deploying a WAR, I know which language gave me the headache!
  10. I don't see many internet hosting services have tomcat installed. Let alone there are incompatibility issues of different versions tomcat. But almost all the internet hosting service providers provide PHP support. I downloaded the PHP installation package. For window NT/2000, the executable version of PHP installation package configured the IIS automatcially while for apache, all you need to do is to make some changes to the httpd.conf file. Even with the war file of java web application file, presonally i think it is a pain. A single line of change in the html file will require to update the war, a zip file which is not as easy to inspect as files in exploded mode. We are just talking about simple vanila jsp war files. We have not got into J2EE yet: EJBs, JDBC configurations, JNDI lookups, and so many evolving web tier packages: struts, spring, JSF.
  11. I don't see many internet hosting services have tomcat installed. Let alone there are incompatibility issues of different versions tomcat.

    Same with PHP.
    But almost all the internet hosting service providers provide PHP support.

    That is a different matter. You were saying that there was a problem due to complexity of configuration.
    I downloaded the PHP installation package.

    I tried that, but as support for my database was not built in to the standard PHP binary, my only option was to recompile the entire PHP source after having to download the source of my database, so PHP could pick up some header files from the database source tree. A messy solution. JDBC (with, of course, a nice ORM above it) is a far better way of working.
    while for apache, all you need to do is to make some changes to the httpd.conf file.
     
    Not if your apache was not originally compiled right.
    Even with the war file of java web application file, presonally i think it is a pain. A single line of change in the html file will require to update the war, a zip file which is not as easy to inspect as files in exploded mode.

    Once it is uploaded, it is exploded by tomcat, and everything is there. If you want to then edit an HTML page, you can. It is then no different from PHP.
    We are just talking about simple vanila jsp war files. We have not got into J2EE yet: EJBs, JDBC configurations, JNDI lookups, and so many evolving web tier packages: struts, spring, JSF.

    But you don't have to use any of those. If you use vanilla JSP files, you are in the same situation as if you are coding with PHP. You can edit a JSP source page just like you can edit a PHP source page. With J2EE you at least have the option of all the higher levels of organisation.

    I am really surprised that anyone could find JSP + a decent set of taglibs any less easy or productive for quick scripting than PHP.
  12. I am really surprised that anyone could find JSP + a
    >decent set of taglibs any less easy or productive for
    >quick scripting than PHP.

    Shameless plug of course, but it is exactly one of the goals for our Coldtags suite: http://www.servletsuite.com/jsp.htm

    And really, with JSP you can code exactly in PHP style if you want. So what is the difference? And Java by the way is not only JSP and servlets.

    But yes, we have to admit that hosting is a real problem for Java, because PHP is almost always included into basic plans. That is why there are 1000’s of references etc.

    Dmitry
  13. PHP vs JSP[ Go to top ]

    For simple projects, PHP is way more productive... if you don't take into account the conversion of PHP to J2EE/.NET later, if you want a truly scalable platform. Truly scalable from not only from an operational, but primarily from a development point of view.

    My personal opinion is that PHP is so attractive, because:

    - hosting is cheap;
    - it's easy to learn;
    - PHP talent is much cheaper;
    - integration w/ MySQL is easy;
    - there are huge libraries of standard functions for virtually everything a web app might need;
    - out of the box PHP has much more stuff already built-in, compared to Tomcat, for example;
    - even if PHP is a scripting language, the libraries are compiled native code, which helps the performance and also makes it much easier to integrate with legacy C/C++;
    - there is so much more sample code online compared to Java;
    - there are numerous advantages of a pure scripting weakly-typed language versus a compiled one;

    There could be other advantages, these are just few of the ones that came to my mind.

    I'm personally a big fan of Java and will always recommend it, but honestly it's hard to convince people that Java is better even for small and mid-size projects.

    I remember a few months ago Sun saying that they will integrate PHP into the Java platform, if I'm not wrong. I personally can't see this happening soon though, because PHP's value is not in it's scripting engine, it is in its libraries. From a pure scripting language, I don't think that PHP is better than JavaScript (especially compared to the new JavaScript versions).

    According to the live edits of JSP - it is possible with Java, yes. I am personally using it and Java IDEs such as JBuilder make it effortless.
  14. PHP vs JSP[ Go to top ]

    there is so much more sample code online compared to Java

    Really? That would seem surprising to me. I did a Google search on sample Java code and then on sample PHP code and I got over 1 million more hits for Java (~1/3 more.) And the Java results seemed a lot less sketchy.

    Not that that is the most reliable metric but...
  15. Oops[ Go to top ]

    It was more but not that much more. Almost 1/2 a million more hits. Sorry.
  16. PHP vs JSP[ Go to top ]

    there is so much more sample code online compared to Java
    Really? That would seem surprising to me. I did a Google search on sample Java code and then on sample PHP code and I got over 1 million more hits for Java (~1/3 more.) And the Java results seemed a lot less sketchy.Not that that is the most reliable metric but...

    I did the same search and picked the first examples that came up. For PHP it was this one:

    <? print("Hello World") ?>

    For Java, it was this:

    import java.text.*;

    public class ChoiceFormatDemo {

       static void displayMessages(Locale currentLocale) {

          System.out.println("currentLocale = " + currentLocale.toString());
          System.out.println();

          ResourceBundle bundle =
             ResourceBundle.getBundle("ChoiceBundle",currentLocale);

          MessageFormat messageForm = new MessageFormat("");
          messageForm.setLocale(currentLocale);

          double[] fileLimits = {0,1,2};

          String [] fileStrings = {
             bundle.getString("noFiles"),
             bundle.getString("oneFile"),
             bundle.getString("multipleFiles")
          };

          ChoiceFormat choiceForm = new ChoiceFormat(fileLimits, fileStrings);

          String pattern = bundle.getString("pattern");
          Format[] formats = {choiceForm, null, NumberFormat.getInstance()};

          messageForm.applyPattern(pattern);
          messageForm.setFormats(formats);

          Object[] messageArguments = {null, "XDisk", null};

          for (int numFiles = 0; numFiles < 4; numFiles++) {
             messageArguments[0] = new Integer(numFiles);
             messageArguments[2] = new Integer(numFiles);
             String result = messageForm.format(messageArguments);
             System.out.println(result);
          }
       }

       static public void main(String[] args) {
          displayMessages(new Locale("en", "US"));
          System.out.println();
          displayMessages(new Locale("fr", "FR"));
       }
    }

    Now you know why PHP is simpler ;-)
  17. PHP vs JSP[ Go to top ]

    Now you know why PHP is simpler ;-)

    I assume you are joking here but I looked at some PHP code examples and it looks like HTML-esque vomit to me. But I'm a Java junky. I'm sure it easy once you get past the aesthetic problems ;)

    Just in case:

    public class HelloWorld
    {
        public static void main(String[] args)
        {
            System.out.println("Hello World");
        }
    }
  18. PHP vs JSP[ Go to top ]

    Actually, if you wanted to compare PHP to JSP, it would be:

    <%="Hello World"%>
  19. PHP vs JSP[ Go to top ]

    Actually, if you wanted to compare PHP to JSP, it would be:<%="Hello World"%>

    ...and could be further simplified as:

    Hello World


    in a WYSIWYG JSP.
  20. PHP vs JSP[ Go to top ]

    Of course I was joking. But the idea behind it was that the perception of there being more useful PHP than Java code samples might stem from the hugely greater breadth of Java. And when I say Java, I mean Java developers at least as much as Java technology. Most PHP code samples are focused on pieces of web user interfaces wheras for Java you will find everything from how to write pooling JCA adapters to OWL inference engines interfacing with RDBMS.
  21. PHP vs JSP[ Go to top ]

    Actually since we can only compare PHP to JSP, the equivalent is:

    <%="Hello World" %>

    Compare that to your PHP ;-).
  22. PHP vs JSP[ Go to top ]

    I did the same search and picked the first examples that came up. For PHP it was this one:

    <? print("Hello World") ?>

    I can't even think of a non offensive reply to that post.

    --b
  23. PHP vs JSP[ Go to top ]

    I did the same search and picked the first examples that came up. For PHP it was this one:<? print("Hello World") ?>
    I can't even think of a non offensive reply to that post.--b

    Then you apparently misunderstood it completely. First, it was anything but serious. It seems nowadays you have to clutter postings with smilies to make that clear, otherwise even gross nonsense is taken seriously and attacked ferociously.

    And second, the point it was meant to support, if any, is that since Java and PHP are used for different purposes, with a small overlap, it should be obvious that the samples you find will also be very different.
  24. PHP is for Simple Minds[ Go to top ]

    <% out.write("Hello World"); %>

    I think the PHP Solution was for a single language, static page that does basically nothing.

    The Java solution is a multi-language, multi-locale solution.

    How would you do a multi-language, multi-locale Hello World in PHP?

    Keep on hackin....
  25. PHP vs JSP[ Go to top ]

    For simple projects, PHP is way more productive... if you don't take into account the conversion of PHP to J2EE/.NET later, if you want a truly scalable platform.

    You make very good points in your post, but I have to admit I am biased, because I have a personal philosophy of software development based on decades of experience of things going wrong....

    "Never assume that any project is simple or small."

    It is my belief that any minor inconvenience involved in using something like JSP, where you have the option of scaling up later, is nothing compared to the state you get into if you are stuck with a pile of scripts in PHP, PERL etc. that need to be ported into something faster and more scalable in a different language. I have seen this situation arise so many times.
  26. PHP vs JSP[ Go to top ]

    For simple projects, PHP is way more productive...

    I don't even know one... I don't even have an application I wrote that is "frozen". They *all* require maintenance and extension...
    - hosting is cheap;

    True. But Java tends to gain marketshares there too...
    - it's easy to learn;

    Well, JSP is *strictly* equivalent, sorry.
    - PHP talent is much cheaper;

    Where have you seen "talent" in writing web pages ??? This must be the most "operational" position in a computer software firm...
    I'd rather see talent in graphics people who make the pages (PHotoshoP rulezzzz :-P)...
    - integration w/ MySQL is easy;

    But impossible with other dbs...
    - there are huge libraries of standard functions for virtually everything a web app might need;

    Java's not poor when it comes to libraries. Check out the various taglibs, etc. There's almost everything there...
    - out of the box PHP has much more stuff already built-in, compared to Tomcat, for example;

    Well, how long will it take for you to find that you need something that's not there ???
    - even if PHP is a scripting language, the libraries are compiled native code, which helps the performance and also makes it much easier to integrate with legacy C/C++;

    Webapps performance is not bound to the nature of the web components, but more to the pooling capabilities (dbs, threads, resources, etc.) that the framework provides.
    Have you got such features in PHP ?

    When it comes to C/C++, well... who would do that ?
    LOL
    JNI allows this anyway, even if IMHO this should never be done this way (memory leaks, etc.). I'd rather CORBA-ize my C code, but that's another question.
    - there is so much more sample code online compared to Java;

    Have you got figures ?
    - there are numerous advantages of a pure scripting weakly-typed language versus a compiled one;

    Sorry, I only see it's far more error-prone, and undebuggable... Script is cool to invoke objects for testing, or manipulate at run time... What's the point in a script that's never updated but interpreted 50 times per second ?
    I'm personally a big fan of Java and will always recommend it, but honestly it's hard to convince people that Java is better even for small and mid-size projects.

    Of course. People who see "small projects" in software development usually have difficulties to understand the benefits of an open and scalable approach... You know, just like the ones who think using WebServices to build distributed systems. All these so-called "simple technologies" work fine until you get out of the "hello world" example, and have a real-life requirement !

    Have fun,

    Remi
  27. PHP vs JSP[ Go to top ]

    <quoteblock>Well, JSP is *strictly* equivalent, sorry.</quoteblock>

    When you say strictly I can only assume you are saying that because java is a typed language and php is loosely typed. This in and of itself does not make them equivelent. Just in the way php handles objects, arrays and functions are completely different. There are a lot of ways you can manipulate data in PHP that are impossible to do in Java.

    The facilities of PHP when it comes to JSP are much different. PHP has many more built in libraries that relate directly to working with web applications. Although you could do everything in JSP, it is infact more bloated and complicated.

    My current job position is as a J2EE developer but also do PHP development on my own time. I believe that anyone who has at least a years+ of experience in both PHP and JSP will tell you that PHP is much more productive and maintainable.

    The naysayers like yourself most likely have only dabbled in php, maybe strung together a few poorly developed screens. There are good design patterns for php as well. That shouldn't come as a surprise though.
  28. PHP vs JSP[ Go to top ]

    When you say strictly I can only assume you are saying that because java is a typed language and php is loosely typed.

    Not at all, this is not what I meant. Sorry if I have problems to express myself, I'm not native english speaker...

    JSP==PhP in the sense that they are "embedded script" technologies. This means a mix of HTML and code.
    When it comes to comparing PHP to Java as a whole, of course it's very different.
    Just in the way php handles objects,

    I don't see a lot of languages that are really OO like Java is. PhP is not in the list. But I admit I may be wrong there : I don't have particular experience in writin OO code in PHP and I don't know the inside of the language itself.
    But I know Java, and the way *it* handles objects is to me very good ! Will be had to de it better IMHO...
    arrays and functions are completely different.

    Well, "arrays" and "functions" are things I almost don't use since years now... Please use nicer and more recent concepts (collections, methods) :-)
    There are a lot of ways you can manipulate data in PHP that are impossible to do in Java.

    Oh yes, sure... Like O/R mappings... Entity beans...
    The facilities of PHP when it comes to JSP are much different. PHP has many more built in libraries that relate directly to working with web applications.

    That kind of statement is once again purely subjective ! Have you recensed all Java/PhP libs and frameworks in the world ???
    Gee, some guys here have really impressive knowledge about statistics, but some of them may come back to the real point (saying the technology) instead of giving us their "feelings" !
    Although you could do everything in JSP, it is infact more bloated and complicated.

    Sure, if you're used to manipulate arrays and functions, I can understand that you have problems with J2EE web frameworks...
    :-P
    But at least you admit you can do *everything*. And to me, it should be enough :-)
    My current job position is as a J2EE developer but also do PHP development on my own time. I believe that anyone who has at least a years+ of experience in both PHP and JSP will tell you that PHP is much more productive and maintainable.

    Believe what you want, fortunately there's many different opinions and that's what makes the world so cool...
    But you have to understand that what you believe is not necessarily true...

    And btw I met *lots* of people who don't plan to switch from Java to PHP for their web apps soon. I personnally have the feeling that the inverse is more frequent :-)
    The naysayers like yourself most likely have only dabbled in php, maybe strung together a few poorly developed screens.

    !
    Gee, is aggressive behaviour required to stay on this forum ??? Have I attacked you in any way ??? If yes I'm very sorry about it, this was not my intention ! I was only enjoying yet another useless but interesting technical discussion... but looks like you're more to have a flame !

    Please don't come and judge me if it's not related to a technical point. You don't have a clue of what I do all day so please keep your remarks for yourself.

    If there was not a few interesting people with something to say here I think I'd unsubscribe soon... Signal/Noise ratio is *really* going down here :-/

    Generic remark : You angry guys, have beer, have a smoke, have a f***, I don't know... Leave your Personnal Home Page for a second and look at the world outside !
    There are good design patterns for php as well. That shouldn't come as a surprise though.

    Finally you reach my point, after your caustic statements !
    Remember, JSP and PHP *strictly* equivalent...
    I was talking about the technology that allows to write "embedded scripts"... Of course you can do MVC & such things with PHP : they are equivalent...

    My problem with PHP is precisely that it's *limited* to an embedded script... And that's why I think the Java "Web Side" is far better since it's the result of an integrated approach.

    Have fun,

    Remi
  29. PHP vs JSP[ Go to top ]

    There are a lot of ways you can manipulate data in PHP that are impossible to do in Java.

    Can you elaborate that? I'm just curious.
  30. PHP vs JSP[ Go to top ]

    For simple projects, PHP is way more productive... if you don't take into account the conversion of PHP to J2EE/.NET later, if you want a truly scalable platform. Truly scalable from not only from an operational, but primarily from a development point of view.My personal opinion is that PHP is so attractive, because:- hosting is cheap;- it's easy to learn;- PHP talent is much cheaper;- integration w/ MySQL is easy;- there are huge libraries of standard functions for virtually everything a web app might need;- out of the box PHP has much more stuff already built-in, compared to Tomcat, for example;- even if PHP is a scripting language, the libraries are compiled native code, which helps the performance and also makes it much easier to integrate with legacy C/C++;- there is so much more sample code online compared to Java;- there are numerous advantages of a pure scripting weakly-typed language versus a compiled one;There could be other advantages, these are just few of the ones that came to my mind.I'm personally a big fan of Java and will always recommend it, but honestly it's hard to convince people that Java is better even for small and mid-size projects.I remember a few months ago Sun saying that they will integrate PHP into the Java platform, if I'm not wrong. I personally can't see this happening soon though, because PHP's value is not in it's scripting engine, it is in its libraries. From a pure scripting language, I don't think that PHP is better than JavaScript (especially compared to the new JavaScript versions).According to the live edits of JSP - it is possible with Java, yes. I am personally using it and Java IDEs such as JBuilder make it effortless.

    Is there a MVC implementation like Struts for PHP or is it like in the old java world where we had a zillion servlets and didn't know which servlet was the backend to a particular jsp/view easily.
  31. PHP vs JSP[ Go to top ]

    http://www.google.lt/search?hl=lt&q=MVC+PHP&btnG=Paie%C5%A1ka&meta=
  32. Re: PHP vs JSP[ Go to top ]

    - hosting is cheap;

    I agree. Sites like sourceforge.net will never be able to afford Java.

    > - it's easy to learn;
    > - PHP talent is much cheaper;

    But the result is much worse. Most of PHP guys don't even know to indent their code correctly, not to mention any knowledge about OOP and MVC.

    > - integration w/ MySQL is easy;

    MySQL is not a serious DB at all...

    > - there are huge libraries of standard functions for virtually everything a web app might need;
    > - out of the box PHP has much more stuff already built-in, compared to Tomcat, for example;

    Ah yes, much easier to use than Java's. But the problem comes if you want to extend their functionality.

    > - even if PHP is a scripting language, the libraries are compiled native code, which helps the performance and also makes it much easier to integrate with legacy C/C++;
    > - there is so much more sample code online compared to Java;

    Agreed.

    > - there are numerous advantages of a pure scripting weakly-typed language versus a compiled one;

    You can also use EL with Map objects :)
  33. Re: PHP vs JSP[ Go to top ]

    Since you bring up my site, we handle 2.5-3 million hits a day on PHP with no problem. There are good and bad Java developers. There are also good and bad PHP developers. But I can promise you that a good developer can create a site with PHP which scales further using less resources every single time. Seriously. PHP has significantly less overhead than Java, if you know what you are doing with Java, it trivial to apply the same principles to PHP. Well, to be honest, you'd have to unlearn the hopelessly overly complex "architectures" that Java developers seem to prefer. But using tools like Smarty make it just as easy to build a maintainable 3 tier webapp in PHP as it is with Java.
  34. Re: PHP vs JSP[ Go to top ]

    Since you bring up my site, we handle 2.5-3 million hits a day on PHP with no problem. There are good and bad Java developers. There are also good and bad PHP developers.

    This all true, but misses something important. PHP is a scripting language for web pages. It may occasionally be used for other things, but that is not what it was designed for. Java is a general purpose high-performance language which includes technologies for scripting web pages. This means that a Java webapp can be seamlessly extended to include other features. Show me a PHP embeddable SQL database, for example. Many tools and libraries for PHP are written in C to give reasonable performance. That can make both deployment and maintainenance a problem. With Java all these things are in the same language and can be deployed as a single WAR file.
    PHP has significantly less overhead than Java


    If this were true, more of PHP would be in PHP, rather than C. It would be nice to be able to support databases other than MySQL without having to recompile the PHP system source for example.
    Well, to be honest, you'd have to unlearn the hopelessly overly complex "architectures" that Java developers seem to prefer.

    You don't have to use any of these architectures if you don't want to. I could say 'use JSP - you don't have to use complex architectures like Smarty or Horde that PHP developers seem to prefer - you can keep it simple'. This would be a silly thing to say, but I hope it makes my point. If you stick to plain JSP with good taglibs (there are plenty of them!) you can write pages very quickly and simply. There is the bonus that behind those web pages you can both write and use very high-performance code without having to drop into C.
  35. Re: PHP vs JSP[ Go to top ]

    Since you bring up my site, we handle 2.5-3 million hits a day on PHP with no problem. There are good and bad Java developers. There are also good and bad PHP developers. But I can promise you that a good developer can create a site with PHP which scales further using less resources every single time. Seriously. PHP has significantly less overhead than Java, if you know what you are doing with Java, it trivial to apply the same principles to PHP. Well, to be honest, you'd have to unlearn the hopelessly overly complex "architectures" that Java developers seem to prefer. But using tools like Smarty make it just as easy to build a maintainable 3 tier webapp in PHP as it is with Java.

    I have worked with PHP since its version 3 was first released, and actually longer than with Java, so I dare say - the above quote is a bullshit. PHP scalability has limits and it is way lower than that of Java's.

    Two quick hints: there is no support for application-scope variables or objects in PHP, everything is per-request so you can not keep any useful information (e.g. permissions and translations) in-memory, leave alone creating a decent in-memory cache. Going back to the database every time you need a piece of information (and database connections are not even poolled) - is a scalability issue. Neither is PHP supporting threads.

    I think noting these aspects, there is not much left to argue about, except religious shout-outs, but if you need some additional objective judgment, I advise to read about the JBoss experience, it is very representative and free of religious blindness:
    http://www.onjava.com/pub/a/onjava/2003/06/04/nukes.html

    As a final remark: I hate and disgust debates like PHP vs Java/J2EE because they are pointless. Both Java and PHP have proven, long ago, that they are very useful in their respective use-cases. I do not see how Java is gonna kill PHP without it stopping being Java or PHP "overtake" Java unless it becomes a clone of Java, just like C# did :P
  36. Re: PHP vs JSP[ Go to top ]

    a final remark: I hate and disgust debates like PHP vs Java/J2EE because they are pointless. Both Java and PHP have proven, long ago, that they are very useful in their respective use-cases. I do not see how Java is gonna kill PHP without it stopping being Java or PHP "overtake" Java unless it becomes a clone of Java, just like C# did :P

    I think such debates do have a point - they help to dispel myths about both Java and PHP. They show that large applications are at least possible in PHP, even if there can be scalability and performance issues, and that JSP can be used in a lightweight and dynamic PHP style. I think there is a lot of ignorance of the capabilities of both approaches.

    I do think that a lot more work could be done to promote JSP+taglibs as an alternative for PHP users. With good tag libraries and use of expression languages, I would suggest that there is little difference in usability. Perhaps parts of these technologies could be re-branded as something like 'JHP', along with a simplified servlet engine.
  37. Re: PHP vs JSP[ Go to top ]

    I have worked with PHP since its version 3 was first released, and actually longer than with Java, so I dare say - the above quote is a bullshit. PHP scalability has limits and it is way lower than that of Java's.
    My PHP experience is not terribly vast, so while I cannot comment on how well it scales for large and very large sites, I dare to say that the below quotes from an experienced PHP developer such as yourself is bullshit at it's best and ignorance at it's finest peak:
    Two quick hints: there is no support for application-scope variables or objects in PHP, everything is per-request so you can not keep any useful information (e.g. permissions and translations) in-memory, leave alone creating a decent in-memory cache.
    How about sessions ? How about memcached ? I'm really not saying that PHP is better than Java; I'm really, really not saying that PHP's answers to most problems are better or at least as good as Java's - as a matter of fact I think the contrary on most accounts, but it really pisses me off to see negligent, lazy and slow-witted programmers bashing about everything they ever managed to get fucked over by due to their own incompetence.
    Going back to the database every time you need a piece of information (and database connections are not even poolled) - is a scalability issue.
    Holy ****, it's in the bloody manual, database connection can be pooled, that's what the god damn persistent connections do.
    I myselft hate and disgust how you bend over for JBoss' brain washing machina while gleefully claiming religious blindlessness (yeah I just made that up). Blessed are the meek.
  38. Re: PHP vs JSP[ Go to top ]

    PHP supports multiple methods of doing application scoped variables. You can use shared memory, or, even better, look at using memcached. We have yet to find a need to use application scope variables though, hitting the database for session data and whatnot has not been a problem. You can dislike the flamewars all you want, but we've done the testing. There is absolutely no way that we could support our 3MM hits a day on the same hardware we're using if we reimplemented in Java. It would require significantly more memory, more and faster processors and page response time drops by nearly half a second.

    You can harp about pooled connections too, but PHP does support them. PHP's persistent connection mechanism is a connection pool. You can set how many max connections and whatnot in php.ini. Calling pconnect pulls a connection from the pool and closing it releases it back to the pool.

    Finally, I've yet to find a need for multithreading on any webapp in any language. If you're looking to do background processing, a much better choice is to spool your data and have a seperate application process the spool.
  39. Re: PHP vs JSP[ Go to top ]

    You can dislike the flamewars all you want, but we've done the testing. There is absolutely no way that we could support our 3MM hits a day on the same hardware we're using if we reimplemented in Java. It would require significantly more memory, more and faster processors and page response time drops by nearly half a second.

    More memory used, probably. More and faster processors and slower response times? Only if you royally screw up. How can you claim "There is absolutely no way .."?

    Look, you're talking about sourceforge, right? And you say:
    We have yet to find a need to use application scope variables though, hitting the database for session data and whatnot has not been a problem.

    So you claim page times would drop by half a second, but you wham on the database to get a session?!? A well-built app would be done writing its last byte back to the client before your app even finishes getting the session data from the database.

    I happen to like PHP just fine for what it is, so I am not harping on it. I just think your claims lack any substantiation. However, I do believe that you could easily construct a test that shows just what you suggested, if you really wanted to. It's just that most people would rather have their apps work well ..

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Shared Memories for J2EE Clusters
  40. you are lead astray[ Go to top ]

    You can of course **** it up in any language, On the other hand, Salesforce and Friendster is proof that you can support heavy load sites both in Java and PHP. But while you really have to make an effort to **** it up with PHP, with Java it comes naturally because the environment and attitude invites you to it. For example when/if you choose a Big J2EE Application Server, you have already have started at a wrong angle.

    The problem with many Java applications that I have downloaded and tried is not so much that they not can support heavy load as that they are already to slow with one user.

    Regards
    Rolf Tollerud
  41. Re: PHP vs JSP[ Go to top ]

    PHP supports multiple methods of doing application scoped variables. You can use shared memory, or, even better, look at using memcached. We have yet to find a need to use application scope variables though, hitting the database for session data and whatnot has not been a problem. You can dislike the flamewars all you want, but we've done the testing. There is absolutely no way that we could support our 3MM hits a day on the same hardware we're using if we reimplemented in Java. It would require significantly more memory, more and faster processors and page response time drops by nearly half a second.You can harp about pooled connections too, but PHP does support them. PHP's persistent connection mechanism is a connection pool. You can set how many max connections and whatnot in php.ini. Calling pconnect pulls a connection from the pool and closing it releases it back to the pool.Finally, I've yet to find a need for multithreading on any webapp in any language. If you're looking to do background processing, a much better choice is to spool your data and have a seperate application process the spool.

    Chris, I think I was very clear in what I meant - PHP does not support application-scope objects or variables. The life-span of PHP objects, as language elements are limited to a HTTP thread. I hope I do not need to explain how this is different from the memcached third-party plugin or shared memory usage.

    Keepeing variables in session is called session-scope variables, not application-scope. One might be surprised, but there is a difference.

    Now regarding the alternatives of improving the performance - the memcached thing and the shared memory.

    First the shared memory - in both the documentation to it on php.net and memcached page it is trashed and criticized and rightly so.

    The thing is - first, it is a direct memory-block-management, in no way - simple applicatin-scope object support, hence absolutely non-elegant to use, second - it seems to be having problems with non-thread-supporting languages, like PHP (see why thread-support is important in a decent language, now?). Read more about it on php.net and memcached site, don't want to garbage the TSS with PHP-related quotes. Not the right place.

    Now, about memcached: First, it has not much to do with PHP. It is a third-party module that is intended for a whole bunch of languages, including Perl - in no way anything particular to PHP. So, my point that PHP does not support application-scope variables, as a language and framework is still absolutely true, alas! Also, LAMP solutions are, mostly, used on shared hostings and I have not seen many that would allow you to have either memcached or use shared memory, hence if you are a software vendor (e.g. CMS vendor) - you would think ten times before using these features, because 80% of your clients won't be even able to install your software. On the contrary, the customers of J2EE products, usually have dedicated servers. Good or bad - that's the reality.

    IMHO, because none of them is a decent language-level support, I am pretty sure as soon as you begin using shared memory and/or memcached for in-process object caching, you will run in huge problems with concurency. Damn, that scary threading and thread-safity, again. Such a pity, PHP totally lacks support for it, eh?

    And last, but not least. You are right about the mysql_pconnect, but
    1) I have seen a lot of PHP code, but quite rarely people using pconnect. Most of the projects use mysql_connect, which is not poolled. Maybe, because of the second argument:
    2) The pconnect has issues, go to php.net/mysq_pconnect and search in user comments, one of them is quite important:

    "Normally you do NOT want to use mysql_pconnect. This function is designed for environments which have a high overhead to connecting to the database. In a typical MySQL / Apache / PHP environment, Apache will create many child processes which lie in idle waiting for a web request to be assigned to them. Each of these child processes will open and hold its own MySQL connection. So if you have a MySQL server which has a limit of 50 connections, but Apache keeps more than 50 child processes running, each of these child processes can hold a connection to your MySQL server, even while they are idle (idle httpd child processes don't lend their MySQL connection to other httpd children, they hold their own). So even if you only have a few pages which actually connect to MySQL on a busy site, you can run out of connections, with all of them not actually being used"

    This is a VERY different situation from solid DB connection pool implementation.

    P.S. Yes, sf.net uses PHP, indeed, but its reliability is really sucky so that just proves the point. It is a shame how often sf.net is down. And they have still not figured-out the login thing the right way. If you are a project admin on sf.net I bet you have gone through the pain of being thrown out all the time, for no good reason. Also, when you re-login - it does not return you to where you were, but gives a lame security error page. I admire sf.net for its ideology and goal, but would never idealize its technical solution. It leaves to hope for much better.
  42. I certainly never have experienced SourceForge.net to go down, TSS was down much more before Howard Lewis got into the picture.

    Anyhow the PHP development process has really accelerated, expect large improvements soon. IBM is not the only one that has noticed. Cedric Beust for example PHP confessions from a Java fiend. On the Tiobe index PHP is the fastest growing language of all (+3.02%), notice that it is more than twice that of Python. Java is down -4.22%.
    http://www.tiobe.com/tpci.htm

    Regards
    Rolf Tollerud
  43. Rolf, you quote the Tiobe index to back your claims. This index is very volatile, so 4% doesn't mean much. If you look at the graph you will notice that Java has regained a lot of ground over the last few months.

    As far as job listing go on dice.com check out
    http://mshiltonj.com/sm/categories/languages/
    Java jobs: 8896
    PHP jobs: 394
    C# jobs: 2533
  44. Last time I checked
    TheServerSide.com is an online community for enterprise Java architects and developers
    .

    Why not the PHP scripties and Tollerudes of the world stick to their forums?
  45. truth about lured of troll[ Go to top ]

    *laugh* - as usual you got a laugh out of me rolf.

    However, I have to say the excessive trolling/ silly news items like this one
    (lets all have an argument about whether screwdrivers or spanners are better shall we???)

    has the potential to lower the quality around here. In the past the thread was often more informative then the article. (if you could dodge the emotive stuff)

    These days the threads are even more full of crap, with less sincere/ valuable contributions.

    rolf: time finish all, and give us a history/ background of the troll. the amount of time you have to spend would tend to suggest some association with TSS or similiar (drum up site traffic with a bit of excitement??)

    TSS needs a new approach to building a subscriber base (religious wars are just sooooo last year).
  46. Hows that for an idea?
    Do a presentation on the motivations and techniques of a professional troll :)
  47. truth about lured of troll[ Go to top ]

    *laugh* - as usual you got a laugh out of me rolf.However, I have to say the excessive trolling/ silly news items like this one(lets all have an argument about whether screwdrivers or spanners are better shall we???)has the potential to lower the quality around here.

    He really does drag down the quality of threads on TSS. They should just limit him to one post per thread.
  48. lured of troll?[ Go to top ]

    One post per thread! That would be cruel as I am only participating in one of 10 -20 subjects. Anyhow I am not saying that "PHP is better", only that Java (and .NET too) are in for some competition. It is wise never to underestimate the opposition.

    Regards
    Rolf Tollerud
    ("drag down the quality of threads") hmf!
  49. lured of troll?[ Go to top ]

    Anyhow I am not saying that "PHP is better", only that Java (and .NET too) are in for some competition.

    There are cases where PHP is better. I can think of a few very easily, including if I'm an ISP and I want to host 200 sites off of a single Linux box and provide some cool features to those sites. What other choices do you have besides PHP/Python/Perl? Java just doesn't cut it for that environment.

    We shouldn't be intimidated that for some things, there are better solutions than Java. In fact, we should be relieved that Java isn't being made into the everything-for-everyone technology, as that will no doubt be the end of its usefulness.

    We should also be looking at these other products and technologies and asking if they can help us do our jobs better, or if there are ways that we can take advantage of them with our existing infrastructures to save time/money. Maybe PHP can be one of the scripting engines supported in Java, for example. Also, if PHP has advantages over JSP, then why don't we study why PHP is better and use that information to improve JSPs?

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Shared Memories for J2EE Clusters
  50. If you are a business person, with a web idea, and limited budget, and maybe a young IT person you can call on then PHP is brilliant.

    The IT person reads a single book, struggles for a day or 2 and they they know everything they need to know - the entire bredth and depth of their technology platform.

    And the job will get done, and your business can start making money.

    Note 1: I have seen this in action.
    Note 2: I have never ever wanted to work in such an environment. Complex data bases and models, data flows, different protocols over the wire, stateful and stateless GUIs in multiple technologies spanning the last 15 years, excel, pdfs, auto execution, massive test beds, millions of rows, massive volume..... wow.

    So PHP is better (sometimes)

    Jonathan
  51. I certainly never have experienced SourceForge.net to go down, TSS was down much more before Howard Lewis got into the picture.

    I take that you have not been browsing SourceForge too much then.
  52. I certainly never have experienced SourceForge.net to go down, TSS was down much more before Howard Lewis got into the picture.
    I take that you have not been browsing SourceForge too much then.
    He spends too much time here, and tons of time googling for "facts" to support his theories and a bunch of time reading literature books coming up with quotes he thinks make sense an d fit the topic.
  53. I certainly never have experienced SourceForge.net to go down

    I've had problems with SF. It's free, so I figure that it's OK that it's often down or slow, and it's definitely gotten much better.
    Anyhow the PHP development process has really accelerated, expect large improvements soon.

    Who needs large improvements? PHP is already very good at what it does. I think the worst possible thing for PHP is to turn it into another Java. OTOH, one of the best possible things for PHP is to formalize ways for it to work with other technologies such as J2EE.

    You watch what IBM does with PHP .. assuming that they do anything, it'll be all around integration with J2EE and other of their infrastructure pieces.
    On the Tiobe index PHP is the fastest growing language of all (+3.02%)

    Actually, ABAP is ;-)

    However, it's great to see that PHP is passing Perl: There is finally hope!

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Shared Memories for J2EE Clusters
  54. Who needs large improvements? PHP is already very good at what it does. I think the worst possible thing for PHP is to turn it into another Java. OTOH, one of the best possible things for PHP is to formalize ways for it to work with other technologies such as J2EE.

    Interestingly this seems to be how a lot of PHP developer feel about the recent changes with PHP: PHP is just fine, don't turn it into Ruby or Java or Python.
    I do like that the Zend guys have started to treat things a bit more formally in defining how classes and objects look and their behaviour in terms of instance lifetime, construction/destruction, pass-by-value|reference and other things; I think not having to guess from hints and user comments how your objects are going to behave is a really good thing.

    I've also seen same unbelievably talented PHP guys wanting to be able to hook up with Java applications running the server, so I'd say that indeed the J2EE integration thing is rather important.
  55. Please, please Cameron, don't add ABAP into this flame! ;)
    I'm afraid of consequences.
  56. Re: PHP vs JSP[ Go to top ]

    Most of PHP guys don't even know to indent their code correctly, not to mention any knowledge about OOP and MVC.
    Are you sure, you have more knowledge about OOP and MVC than most PHP guys ?
  57. Re: PHP vs JSP[ Go to top ]

    Most of PHP guys don't even know to indent their code correctly, not to mention any knowledge about OOP and MVC.
    Are you sure, you have more knowledge about OOP and MVC than most PHP guys ?
    I'm sure most of J2EE developers do.
  58. Re: PHP vs JSP[ Go to top ]

    MySQL is not a serious DB at all...


    What do you base that on, scientific or theory?

    Google uses MySQL AFAIK (and you can learn more:
    http://www.mysqluc.com/pub/w/35/tutorials.html )

    Theory is O/R.
    Show me the track recrod of sites that can reproduce. In PHP... you don't code, you get an application and customize it. (and it's E/R based - so data access is faster)

    .V
  59. Re: PHP vs JSP[ Go to top ]

    MySQL is not a serious DB at all...
    What do you base that on, scientific or theory?
    The fact that MySQL still is playing catch-up when it comes to supporting features that most serious databases has supported for years now (if not decades ...), perhaps? But MySQL is getting there, and if you don't need more features than MySQL supplies, I suppose you're just fine...
    In PHP... you don't code, you get an application and customize it.
    Kinda hard to beat a platform where every application is already written for you, I have to admit that ...
  60. I am really surprised that anyone could find JSP + a decent set of taglibs any less easy or productive for quick scripting than PHP.

    I am as well...why are we using a new language to write scripts again? I already have that with JSP, and my developers don't need to learn another language. It was tough enough to get them through the learning curve from COBOL.
  61. I recently spent far too much time dealing with the installation of a major PHP application, which involved complicated configurations and compilations of apache and PHP

    Agree, this is my experience too. Remember PHP is short for Personal Home Page.
  62. I recently spent far too much time dealing with the installation of a major PHP application, which involved complicated configurations and compilations of apache and PHP
    Agree, this is my experience too. Remember PHP is short for Personal Home Page.

    I have started believing, people on this discussion saying "... donot understand any thing about server side programming ..."

    I am surprised that you took time to install a php and apache setup, but that aside, PHP has its limitations but is a wonder ful langage to work with. At the same time JAVA is a great tool to build Enterprise applications, each has its place and it is just sad to read people saying "with php so and so concept is wrong, and this problem doesnot exist with JAVA and stuff like that" For me PHP is just an equivalent of JSP or even say ASP.

    As a langauge just defines how html and a scripting language can work together to get some dynamic content out.

    But using it just like that is one of the worst ways to build WEB applications be it in JAVA, PHP or ASP. A good think with JAVA has been focus on Design patterns, a good amount of fuss about frameworks and stuff like that which is good.

    At the same time PHP has a great set of tools and libraries that enable use to build secure, easily maintainable and scalable solutions. A tool like smarty, with PEAR and basic OO concepts with PHP can help build a very scalable, easily maintainable, skinnable and database agnostic (to some extent) web/enterprise application very easily.

    On the PHP side there have been some attempts to make it look like JAVA, (personally, I feel that is not a requried step for PHP's road map, making it JAVA's poor cousin). JAVA has its place and PHP has its place.

    It is more important to leverage strengths of both, and use XML to glue applications/components where required.

    For the past 4 years we have been working with JAVA and PHP for different needs, see PHP grow, just as we saw JAVA grow, and feel a little disheartened with this kind of a discussion in one of the best forums on Internet for Server side programming!!!

    The blame is not in the langauge/platform but in how people care to develop applications using the tools.

    Sarath.
    http://www.quadone.com
  63. Somebody should have pointed this out earlier -
    PHP is not a language - it is a large set of assorted libraries joined up with a sprinkling of syntax (much like PERL [Pathologically Eclectic Rubbish Lister]). The structure recently introduced objects and PEAR, is a great improvement, but is poorly integrated and too little too late.
    (I've fixed enough PHP scripts to have a healthy dislike for the language)

    Sometimes scripting is easier, but you can still use a real language eg Python or Ruby.
  64. The official press release[ Go to top ]

    http://www-306.ibm.com/software/data/info/zendcore/pr.html

    IBM goes where the demand is:

    quote: As a result of this announcement IBM becomes the first vendor in the industry to support both major Web development languages (PHP and Java), with an integrated database solution. According to the research firm Netcraft, PHP currently accounts for more than 40 percent of the overall Web scripting language market.

    The announced solution will be interoperable with Cloudscape and DB2.

    All in all this will probably play well for java as Cloudscape requires java installed.

    In the end, IBM hopes the solution to become a DB2 distribution vehicle, nothing more and nothing less.
  65. That's funny[ Go to top ]

    I find it funny that IBM think J2EE devlopement is too complicated. J2EE is simple, it's the vendors like IBM that make the platform difficult to cope with. Look at their portal and app servers totally overkill for 99.999% of applications and too damn expensive.
  66. That's funny[ Go to top ]

    I find it funny that IBM think J2EE devlopement is too complicated.

    Lets see how the 2 stack up! Here is a list of PHP Applications and some of them have thousands of reference sites.
    http://www.opensourcecms.com/index.php?option=com_wrapper&Itemid=139

    I would *LOVE* to say that Java is supperior emotionaly, but intelecualy, please show me a list of Java prouduction sits or a list of Java applications.

    tia,
    .V
  67. That's funny[ Go to top ]

    I'm not sure what you're getting at. First, I never said anything bad about PHP. My comment was pointing out the hypocrisy of IBM in this particular case. Second, You can find a list of java desktop apps here http://www.javadesktop.org/tsc/sightings/S20.html and this is by no means an complete list. Third, I've personally written almost a hundred J2EE apps that are in production at various companies for interanl and external use. I also have some experience with PHP and find it nice for many things.
  68. PHP vs Java[ Go to top ]

    Here is a list of PHP Applications and some of them have thousands of reference sites.http://www.opensourcecms.com/index.php?option=com_wrapper&Itemid=139I would *LOVE* to say that Java is supperior emotionaly, but intelecualy, please show me a list of Java prouduction sits or a list of Java applications.

    repost - can we drop theory and look at track record?
    .V
  69. That's funny[ Go to top ]

    Lets see how the 2 stack up! Here is a list of PHP Applications and some of them have thousands of reference sites.http://www.opensourcecms.com/index.php?option=com_wrapper&Itemid=139I would *LOVE* to say that Java is supperior emotionaly, but intelecualy, please show me a list of Java prouduction sits or a list of Java applications.

    What you have posted is a list of various cms systems. Some are written in Java, so I guess you've just answered your own question.

    In a sense, you have a point; The major part of opensource cms and portal systems are written in PHP. After all, that's what it was designed to do well. And now that PHP 5 has been released, many of the nice Java-web-frameworks are being ported over to PHP. That is of course easier now that PHP looks syntactiacally more like a mature programming language, like Java. So, yes, PHP is getting better and better while catching up with Java.

    The fact that PHP is powering so many sites around is more a question of price and availability, rather than suitability. PHP is, and has been for years, bundled with every Linux distribution available, and Java has not. PHP is free, while Java requires a web container, and possibly an ejb container (unfortunately PHP doesn't offer a component architecture like ejb just yet ...) to be useful. So, what is the most likely choice when one just want to get a site up and running as fast as possible?

    Finally, let me ask you a question back; can you provide a list of banks, hospitals and airlines running their backend system on PHP?
  70. re: that's funny[ Go to top ]

    Not really funny. They do use struts, after all. no wonder php looks nice...
    LOL...
  71. That's funny[ Go to top ]

    I find it funny that IBM think J2EE development is too complicated.

    If you reread the quote, it came from the unnamed "analyst", not IBM.
  72. sorry but ...[ Go to top ]

    Sorry, but many parts of J2EE are not simple. For example EJB. The parts that are simple (Servlets/WAR) have been wildly successful, whereas the parts at the other end of the complexity spectrum have burned many. Part of the problem has been the misdirection of heavyweight techniques to simple problems (re: "how to build your homepage with EJB's, DAO, etc.").

    The future of Java depends on creating simple, robust tools, that are developer driven (be they OS or commercial). Hibernate is a great example, Kodo JDO is a great example, JAXB from Sun is a great example. All these tools/standards SOLVED PROBLEMS for developers. Hence they met adoption.

    I think the cummunity has taken these lessons to heart. But we'll see. The JDO Reconsideration Ballot is still pending. If the Java Executive Committee itself can't see the value in JDO and other lightweight standards, who do we have to blame for proliferation of PHP and other alternatives to Java?

    -geoff
  73. sorry but ...[ Go to top ]

    The future of Java depends on creating simple, robust tools, that are developer driven (be they OS or commercial). Hibernate is a great example, Kodo JDO is a great example, JAXB from Sun is a great example. All these tools/standards SOLVED PROBLEMS for developers. Hence they met adoption.I think the cummunity has taken these lessons to heart. But we'll see. The JDO Reconsideration Ballot is still pending. If the Java Executive Committee itself can't see the value in JDO and other lightweight standards, who do we have to blame for proliferation of PHP and other alternatives to Java?-geoff

    Part of the problem with now with Java is that there are so many tools that no one can keep up with all of them. A lot of great ideas flounder because the writer is unknown or improves on something that everyone has already adopted.

    JAXB, BTW seems like a great tool but it's a virus that turns your object model into smelly goo.
  74. the playtime is over, they blew it[ Go to top ]

    For many years the Java developers have played. The play is called "the Computer Scientists". Talked endlessly of patterns, "Gang of four", Fowler and you name it, while the failure rate for J2EE projects is somewhat around 70-80%. "Organize inter-service transfers according to use cases from known domain objects into a coarse-grained Composite".

    Well the playtime is over it is quite natural. The only thing that puzzles me is why it took so long.

    Regards
    Rolf Tollerud
  75. the playtime is over, they blew it[ Go to top ]

    while the failure rate for J2EE projects is somewhat around 70-80%.

    You are confusing J2EE with EJBs again, aren't you? A failure rate of 70-80% for JSP/Servlet projects is obviously nonsense.

    Even if it is EJBs alone, I am assuming that this statistic is based by your usual quality of evidence.
  76. Rolf, please, quit being the devil's advocate, trying to find an advantage for .NET in every single piece of news! In this case, if PHP is the winner, then both Java and .NET should be considered loosers!

    Pleae, understand that this is a Java community (it's TheServerSide.com, not TheServerSide.net) and unsuccessfully trying for so long to proove that it's a community of loosers, is a complete waste of time! The fact that Java (as any modern technology) has challenges, does not make ASP.NET is a better choice!

    Sincerely, I like Microsoft as the most powerful software company even, I like (and prefer in most cases) Windows and Pocket PC, I think MS Office is the best office product, but that's it! Also, years ago I used to like Visual Studio, it was a modern IDE back then, but now (and even with 2005), it is still years behind the average Java IDE!

    My personal experience with web technologies is as follows: I've done a lot of ASP (5 years), some ASP.NET (2 years), some JSP/JSTL/Struts (3 years), and also some PHP (1 year). Base on my experience, my practical (not emotional!) favorite is Java and especially recently with JSF!

    If you take productivity as the whole process from the specs to the production code plus some near-term future tweaks, then definitely JSP is better than even PHP. If you care about quick out-of-the-door proof-of-concept, PHP could be better... but my personal opinion is that even in production, PHP will still be a POC quality. As I said earlier, Java talent is more expensive, but for a reason.
  77. I guess it's not just an unnamed analyst's opinion, take a look at zend.com.

    Afterall, you can't look at the case as java vs php challenge, as the choice strongly depends on the project's requirements. As mentioned above, for small and mid-sized standalone webapps using php may be beneficial. Php is fast and easy indeed, however it encourages writing messy code. Java with its OO nature is surely better to maintain, scale etc, which make it fit large and complex apps.
  78. Enough is enough![ Go to top ]

    In the latest project which is kind of "client-Server on steroids" (C# Xamlon click-once clients calling SOA servers) we already have given up either Java or .NET at the server.

    For the communications layer a very simple XMl protocol was chosen. The alternative was 3-4 serializations back and forth between objects and XML during transport..
    "The bigger question is whether it is cost-effective to do this translation from the non-OO world to the OO world.."

    "The answer is different for different services. Many services merely act as gateways to a database or other data store. They do little actual processing or work on the data. In such cases it is hard to imagine justifying translation of the data into objects, when those objects do little to no actual work.."

    "A service/component is merely a group of atomic and stateless procedures or methods."
    Read the article"The intersection of Services".

    Since the server-side was so simple it we adopted KISS and did it in PHP. In the backside of our minds was also that it will make the deployment easier and the (potential) market wider.

    But don't worry. I am sure that with just a tiny little stretch of imagination it is always possible to find reasons to do as many translations of data into objects (and back) as one could ever desire.

    Regards
    Rolf Tollerud
  79. the playtime is over, they blew it[ Go to top ]

    For many years the Java developers have played. The play is called "the Computer Scientists".

    With you around, it always felt more like whack-a-mole.

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Shared Memories for J2EE Clusters
  80. You know something Cameron? We have opinions very different you and I. Lets call yours A and mine B.

    Now what has happened during the three years here in TSS is that "the common opinion" has moved considerable in the B direction. Please don't deny it, spare me the work of digging up old saying and posts. (But it is comfortable to know that everything is there).

    So the situation is, I know that you know. I also know that you know what I know! So what’s the use?

    And the game is not over. By that I mean that you (the J2EE/EJB guys) are not humiliated enough yet! :)

    With utmost respect
    Rolf Tollerud
  81. You know something Cameron? We have opinions very different you and I. Lets call yours A and mine B.

    Now what has happened during the three years here in TSS is that "the common opinion" has moved considerable in the B direction.

    Rolf, you've already exhibited your predilection for ex post facto predictions. Perhaps you can look "ahead" and tell me who is going to win the previous Superbowl?
    Please don't deny it, spare me the work of digging up old saying and posts.

    In that case, I deny it. Please waste some time digging up old sayings and posts.
    So the situation is, I know that you know. I also know that you know what I know!

    By having occasionally responded to you, I unfortunately am probably largely responsible for your delusions of grandeur. I apologize to the rest of the users of this site for my tragic mistake.
    And the game is not over. By that I mean that you (the J2EE/EJB guys) are not humiliated enough yet!

    I am glad to see that you are capable of admitting what a small, small man you are. However, you are only hurting yourself.

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Shared Memories for J2EE Clusters
  82. the Pearl of the day[ Go to top ]

    I knew that you would,

    a) Admit nothing
    b) Revert to Ad Hominem attacks

    You naturally have no idea of the importance of cricket, not being British.

    Therefore I thought that I rather, instead of wasting time on a mountainous dug-up effort instead now and then could present "Pearls from Cameron’s wisdom", a column so to speak.

    In August 20, 2001 Cameron said
    (in response to Anecdotal evidence suggests that many developers remain skeptical about the value of Entity Beans)

    "I am not sure where you collect anecdotal evidence, but I can assure you that it is fundamentally invalid. Most of the companies that I have worked or spoken with over the past year have based their transactional data models in their new J2EE applications entirely on entity EJBs. These include, for example, most of the major financial institutions on the east coast and several of the largest in Europe."
    The situation today:
    When Rod Johson in his the talk 'J2EE without EJB' at Javapolis asked how many people were still using entity beans only a few people (out of maybe 300 persons) raised their hands; so clearly people have left entity beans behind..
    http://www.theserverside.com/news/thread.tss?thread_id=28995#149925
    Alexander Jerusalem
    Now we have server side applications that behave like 1992 desktop applications.
    http://www.theserverside.com/news/thread.tss?thread_id=24823#116183
    Vic Cekvenich,
    What makes you think that I would ever want to help an EJB JCP? How many people are burnt by EJB? How much prepetuated lies? "EJB is scalleable". It's almost expected to lie in J2EE world. That whole group had put quite hurting on my work for many years, helping them would be like “Stockholm syndrome”.
    http://www.theserverside.com/news/thread.tss?thread_id=31712#156248

    (more to follow, like Cameron views on C#.NET, his favorite companies Sun and Weblogic, fake microbenchmarks, personal attacks on people, etc etc, all in his nausatious Elderly Statesman tyle)

    Do you think I can get the column syndicated? (as always I am interested in earning money :)

    My very best regards
    Rolf Tollerud
  83. Feature Suggestion[ Go to top ]

    I'd like a feature to mark a person noisy (Rolf). (and this article)

    I come here to read about technology (specifically Java), not about C# or PHP. (C# is ok, most php is write only code)

    I think they put up www.thserverside.net for both of C#'s users (joke)... so Rolf please take your flaming over there.
    (you need to come to a Jave forum to have people see it)

    -jp
  84. Wow, guys, this is the worst flame war I've seen on TSS yet.

    Firstly, you're not even comparing apples with apples. PHP is targeted at developing web applications. Java on the otherhand has many different flavors, J2SE, J2ME and J2EE, each targeting a different problem space. When complimented by many other APIs from other sources you have a very flexible tool to solve many different problems.

    To compare PHP with Java is as fruitless as comparing Cobol to machine code.

    Pick the right tool(s) for the job at hand and exploit it's features to solve the problem as best as you can.
  85. Pick the most appropriate tool for the task[ Go to top ]

    Wow, guys, this is the worst flame war I've seen on TSS yet.Firstly, you're not even comparing apples with apples. PHP is targeted at developing web applications. Java on the otherhand has many different flavors, J2SE, J2ME and J2EE, each targeting a different problem space. When complimented by many other APIs from other sources you have a very flexible tool to solve many different problems.To compare PHP with Java is as fruitless as comparing Cobol to machine code.Pick the right tool(s) for the job at hand and exploit it's features to solve the problem as best as you can.

    And I build apps that have a "web interface". They, may or may not have another interface (user or otherwise) so that makes using PHP, or any other language like it, difficult to do.

    I kills me when people say "web app". Somethings are web apps (meaning they cannot be usefully anything else).
  86. the Pearl of the day[ Go to top ]

    The situation today:
    When Rod Johson in his the talk 'J2EE without EJB' at Javapolis asked how many people were still using entity beans only a few people (out of maybe 300 persons) raised their hands; so clearly people have left entity beans behind..

    So your only evidence is that at a talk titled "J2EE without EJB" most of those attending were doing J2EE without EJB. Fancy that! ;)

    You can find opinions to say whatever you want. Give us facts.

    Reminds me of the time you took on the guys at JavaLobby in a benchmarking war (C# vs Java) and lost horribly.
  87. the playtime is over, they blew it[ Go to top ]

    For many years the Java developers have played. The play is called "the Computer Scientists". Talked endlessly of patterns, "Gang of four", Fowler and you name it, while the failure rate for J2EE projects is somewhat around 70-80%. ....RegardsRolf Tollerud

    I have watched a group of ex. VB/ASP junkies work with .NET for a year now. The website uptime is now down the toilet - restarts every day sometimes hourly, code littered around and copied like in the good old ASP days, development times have ballooned out of control, constant .NET related quirks and problems to the extent that nobody knows how to fix them.

    This group could use a couple of computer scientists, badly! It just makes me sigh. I was working with them briefly, and we did not even have a common base level language to exchange ideas. It was hard to even help the solve problems, because you get flooded with basic level questions that should be common knowledge had you gone to any kind of center of higher education - not just your 8 month programming course that would make you an 'expert'.

    Not everything was their fault. I mean come on. As I started programming with .NET it quickly became apparent that the .NET libraries were riddled with bugs. I don't know how much time I have wasted on idiotic memory leaks, caused by .NET collections etc. etc. As far as I am concerned, .NET is only a couple of months old now, since the v1.1 service pack came out. Before that, the software was totally in beta stage.

    Compared to the group I just described, the MS people are total computer scientists who failed to deliver something understandable to their user base.
  88. confused VB/ASP junkies[ Go to top ]

    Tero,

    "This group could use a couple of computer scientists, badly!"

    No. If what you tell is true they only need a "Vic Cekvenich type of guy".

    Regards
    Rold Tollerud
  89. the playtime is over, they blew it[ Go to top ]

    Please share some acceptable links/references which says the failure rate is 70-80% for J2EE projects.

    It is good to have at least one solid statistical proof before we declare anything quantitatively about the failures.
  90. be my quest[ Go to top ]

    To date, around 70 percent of initial Java implementations have been unsuccessful, according to new research from Gartner Group.


    http://www.zdnet.com.au/news/business/0,39023166,20269968,00.htm

    Please don't take any comfort from "In all likelihood, the failure rate for early implementations of .NET systems will be similar". That is just a political statement, while the Java number come from research.

    In any case, the PHP failure rate is nowhere near these numbers.

    Regards
    Rolf Tollerud
  91. be my quest[ Go to top ]

    To date, around 70 percent of initial Java implementations have been unsuccessful, according to new research from Gartner Group.
    http://www.zdnet.com.au/news/business/0,39023166,20269968,00.htmPlease don't take any comfort from "In all likelihood, the failure rate for early implementations of .NET systems will be similar". That is just a political statement, while the Java number come from research.In any case, the PHP failure rate is nowhere near these numbers.RegardsRolf Tollerud

    Statistician Steve here again. As usual Rolf, you hunt around for reports that back up your point of view, no matter how biased or out of date.

    This so-called 'new research' was from 2002. You have no evidence for the PHP failure rate, or the .Net failure rate, and so you can't meaningfully compare them to Java - you are just making political statements yourself. There is a good reason why initial (and not necessarily final) implementations of Java fail - they are being used for larger projects than PHP. PHP is a web page scripting language. Adding some basic form functionality to web pages (a typical PHP project) is nothing compared to developing a business application. To compare like-with-like you would have to compare PHP projects with JSP projects. Any other comparison is meaningless.

    Oh, and could you please quote statistics that are up-to-date before labelling them 'new'?
  92. be my quest[ Go to top ]

    To date, around 70 percent of initial Java implementations have been unsuccessful, according to new research from Gartner Group.
    http://www.zdnet.com.au/news/business/0,39023166,20269968,00.htm
    The title of the story is:
    Java and .Net both a disaster: research

    This is nothing new about failure in IT projects. This is more a reflection on project mangement than anything else.
    http://www.it-cortex.com/Stat_Failure_Rate.htm
  93. php is better[ Go to top ]

    its life cycle for developer is essentially shorter as it is a scripting language and there is no container to care about and no deployment phase and best of all there is no incompatabilities as there is only one php distro vendor(and the community is forcind unfragmentation) instead of different j2ee vendors ,each trying to b unique by defining features not defined in the specs(ok this is the fault of the jcp EGs as they donot cover every aspect in the specs,leaving enough room for incompatible implementations and unportable application)
    BUT BUT BUT
    php is lacking (in present) transaction support and implicit security model and good ORM tool and............
    SO it is not the time to say that php will replace java
    but it will happen someday if the jcp doesnot change its loooooooooooooong cycle and its fool strategy
  94. php is better?[ Go to top ]

    its life cycle for developer is essentially shorter as it is a scripting language and there is no container to care about and no deployment phase and best of all there is no incompatabilities as there is only one php distro vendor(and the community is forcind unfragmentation) instead of different j2ee vendors ,each trying to b unique by defining features not defined in the specs(ok this is the fault of the jcp EGs as they donot cover every aspect in the specs,leaving enough room for incompatible implementations and unportable application)
    What I dislike in PHP (during my 3 years of PHP programming experience) is following:
    1. No container - you ara up to your self with middleware apsects of your system. You need to care about many many staff container provides you with. Simplicity of the PHP goes along with SLOOOOOOOOOOW perfomance, no caching, no instance pooling, etc.
    2. No deployment phase. You will never really know wherever your PHP APP is deployed fully, wherever you haven't forgotten on neither of the 1000 PHP files nor of the 10'000 images and html pages.
    3. No testing at all - no comments on this one.
    4. One PHP vendor, if it doesn't fit - you are stuck.

    But the most important one (which is changing now) is a language itself - NO PRESENCE OF MODULARITY CONTRUCTS, which make almost impossible support and evolution of the complex system.
    And offcourse
    "php is lacking (in present) transaction support and implicit security model and good ORM tool and".
    I think as soon as Java will be called PHP I'll use it.




    BUT BUT BUTphp is lacking (in present) transaction support and implicit security model and good ORM tool and............SO it is not the time to say that php will replace javabut it will happen someday if the jcp doesnot change its loooooooooooooong cycle and its fool strategy
  95. Java may need more time for configurations and so, yet there's nothing that can replace it's ability to intergrate scattered parts in single boil.
  96. Unnamed executive, not analyst[ Go to top ]

    Not that it matters a whole lot, but unnamed person being quoted is an executive, who likely has a different set of priorities from your typical analyst.
  97. definitive answer[ Go to top ]

    This post was a flame-war setup (a redundant one at that) if I've ever seen one :)

    Personally, I like coding in Java more than PHP. Java is fun.
    That is all that matters ;)
  98. The postings on TSS are beginning to suck. Named developer makes jab at TSS - Me
  99. Who would pay an Un-named analyst.
    if some one could be an analyst and could back what he/she
    says then his/her analysis will contain some identity.
    IBM disillusioned from java is the most unnatural thing which an Analyst (from un-named kind) could talk about. specially in the time which IBM update all its development/deployment stuffs recently.
  100. Why do people think Java is a language for writing web pages? Java is a multipurpose language. JSP is a scripting languange that meshes with Java. It's hardly all that Java does.

    I have used Java for years and very little of what I write has anything to do with web sites.

    I can't claim to be an expert on PHP but from what I gather it's mainly for web development. I don't see how it's a replacement for Java. It seems to me that it would be a candidate for integration with Java like Jython.
  101. I can't understand why anyone would use PHP in a medium to large site. You can't even write a database connection pool in PHP (last time I checked it was not possible).

    Anyone who claim's it's easier to work with PHP than JSP obviously has never worked with JSP. The age of script kiddies is slowly dying (I honestly hope so), go learn concepts and design. They will apply no matter what language/platform you use.
  102. I can't understand why anyone would use PHP in a medium to large site. You can't even write a database connection pool in PHP (last time I checked it was not possible).Anyone who claim's it's easier to work with PHP than JSP obviously has never worked with JSP. The age of script kiddies is slowly dying (I honestly hope so), go learn concepts and design. They will apply no matter what language/platform you use.
    http://www.felixgers.de/teaching/php/db_persistent.html
  103. I can't understand why anyone would use PHP in a medium to large site. You can't even write a database connection pool in PHP (last time I checked it was not possible).Anyone who claim's it's easier to work with PHP than JSP obviously has never worked with JSP. The age of script kiddies is slowly dying (I honestly hope so), go learn concepts and design. They will apply no matter what language/platform you use.
    http://www.felixgers.de/teaching/php/db_persistent.html

    Uh, that is a PHP function, implemented in the C programming language, just like the non-persistent PHP db connections. Most connection pools in the java world is written in java. You have exactly one connection pool implementation to use when running PHP.
  104. I found a interesting site
    "News, info, tutorial, resource from the Flash and PHP world". With PHP 5, I think we can not dismiss PHP and Flash sail up as a important contender to ClickOnce/Java Webstart in the RIA space.

    Regards
    Rolf Tollerud
  105. You can't even write a database connection pool in PHP (last time I checked it was not possible).

    Last time I checked, about 5 years ago, PHP had persistent database connections to the back-end.

    Listen, you don't know what the **** you're talking about, why even bother? You may be spot on with some remarks, but for ****'s sake people don't just splat out assuptions and hearsay just because you know you're basically right.

    JSP is really fucked up and in no way a better choice for pages over PHP, as it comes with all the Java "bloat" like checked exceptions, strong typing, large framework.

    On the other hand, the same "bloat" is what makes Java far more suitable for like, the rest of code that's not presentation layer. I'd always write back-ends in Java, .NET or C++ but hardly want to their web-tier incarnation (JSP, ASPX) for front-end.
    go learn concepts and design. They will apply no matter what language/platform you use.
    That is to say they can apply to PHP too, isn't that right Mr. SelffuckingrighteousthatlikesJSPandthinksPHPpeoplearescriptkiddies ?
    What a superbly arrogant attitude combined with such graceful ignorance, truly /. style.

    Also, I have worked with JSP and have also worked with PHP and I'm saying yes PHP is easier and I have obviously not never worked with JSP. Bloody hell.
  106. You can't even write a database connection pool in PHP (last time I checked it was not possible).
    Last time I checked, about 5 years ago, PHP had persistent database connections...

    Before you begin to mouth off at me, learn something. A database connection pool and persistent database connections are totally two different things, they even have different names moron.

    A database connection pool is the concept of opening X number of database connection and keeping them open. Then re-using them as nessesary. The pool may also grow to Y number of connections, but no more than Y. Thereby giving you performance with flexibility.

    Persistent database connections are just connections that are never closed. There is no cap, so a PHP application can take a database server down my usurping all available connections. From what I understand it is impossible to create a database connection pool due to the process like nature of apache (instead of the thread-like nature of JSP/Servlets).

    Please read before you speak, otherwise you come off sounding like a complete idiot.

    http://www.felixgers.de/teaching/php/db_persistent.html
    "Drawbacks of Persistent Connections
    If you are using a database with connection limits that are exceeded by persistent child connections. If your database has a limit of 16 simultaneous connections, and in the course of a busy server session, 17 child threads attempt to connect, one will not be able to....

    There are a couple of additional caveats to keep in mind when using persistent connections. One is that when using table locking on a persistent connection, if the script for whatever reason cannot release the lock, then subsequent scripts using the same connection will block indefinitely and may require that you either restart the httpd server or the database server."
  107. Both have drawbacks, but probably information about PHP is more honest, less hype and marketing stuff. There is no technical problems with PHP and database integration, you just need to configure it in the right way. It is very good stuff for web development ( "easy to learn" for web masters). There are a lot frameworks for PHP too (including MVC and ORM). But I am afraid IBM can fail to integrate it with mainframe stuff for political reasons (PHP is not known as "enterprise framework").
  108. A database connection pool and persistent database connections are totally two different things, they even have different names moron.
    Actually they're not totally different things, moron. A DB pool is a pool of connections, that means you get connections from it, put those back and that's it. It may have lwm and hwm, guaranteed availability, validate-before-use, transparent reconnect, server fail-over, several policies to deal with out-of-resources situations.
    Now PHP's persistent connection mechanism is a connection po-fucking-ol my dear because you get connections from it, put those back, it does not close them for you and it just has hwm. Indeed it's very basic, but it is a db pool.
    There is no cap, so a PHP application can take a database server down my usurping all available connections.
    That's bollox. RTFM if you really must comment - of course you can very well bloody limit the number of database connections created. Do you just make that up as you type ? See php.ini and count max_persistent.
    From what I understand it is impossible to create a database connection pool due to the process like nature of apache (instead of the thread-like nature of JSP/Servlets).

    Umm, like, see the above. Also shared code between all processes is obviously possible. Also non-shared code between all processes that is still able to act as a single entity is possible.
    Please read before you speak, otherwise you come off sounding like a complete idiot.
    Right.
  109. A database connection pool and persistent database connections are totally two different things, they even have different names moron.
    Actually they're not totally different things, moron. A DB pool is a pool of connections, that means you get connections from it, put those back and that's it. It may have lwm and hwm, guaranteed availability, validate-before-use, transparent reconnect, server fail-over, several policies to deal with out-of-resources situations.Now PHP's persistent connection mechanism is a connection po-fucking-ol my dear because you get connections from it, put those back, it does not close them for you and it just has hwm. Indeed it's very basic, but it is a db pool.

    Each apache child process keeps and reuse its own db connection until the process dies. Strictly speaking, there is no db connection pool, only an apache process pool.
    There is no cap, so a PHP application can take a database server down my usurping all available connections.
    That's bollox. RTFM if you really must comment - of course you can very well bloody limit the number of database connections created. Do you just make that up as you type ? See php.ini and count max_persistent.
    From what I understand it is impossible to create a database connection pool due to the process like nature of apache (instead of the thread-like nature of JSP/Servlets).
    Umm, like, see the above. Also shared code between all processes is obviously possible. Also non-shared code between all processes that is still able to act as a single entity is possible.

    Impossible is nothing. But in reality, php persistent connection will be kept open for a process, and it will not allow connections to be shared between different processes. It's not as scalable as true db connection pooling. Plain old ora_logon() with SQLRelay is a better solution.
  110. Probably most PHP applications do not use persistent connections (connect once per request), it is more scalable than pools with backround threads, fake query execution,reconnection and abandon stuff. LAMP doe's not has problems with pools and caches, because smart PHP developers know how to develop scalable sofware without workarounds. I respect "JAVA camp" arogance and ignorance, but it is a good idea to learn competitor tools too.
  111. Probably most PHP applications do not use persistent connections (connect once per request), it is more scalable than pools ..

    I don't understand. Please explain how establishing a secure connection to a database is more expensive than doing nothing (i.e. getting a pre-existing connection out of a pool).
    LAMP doe's not has problems with pools and caches, because smart PHP developers know how to develop scalable sofware without workarounds.

    Pools and caches are not workarounds; they are logical ways to increase scalability by reducing the load on the network and on other servers.

    If you disagree, then I suggest that you turn off the L0, L1 and L2 caches in your CPU and see how your computing experience improves "without workarounds" ;-)

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Shared Memories for J2EE Clusters
  112. Just try it, install a few hundred PHP applications without workarounds on webserver A and a few hundred JAVA applications with pools and caches on webserver B. I hope you will see the reason for PHP hosting price.
  113. Just try it, install a few hundred PHP applications without workarounds on webserver A and a few hundred JAVA applications with pools and caches on webserver B. I hope you will see the reason for PHP hosting price.

    Oh, I completely agree with that. In fact, I said as much above.

    When you said "scalability", I thought you meant numbers of servers, numbers of concurrent users, etc., for a single application. FWIW I just got off the phone with a customer with a large Coherence cluster that handles significantly more hits than the cited sourceforge numbers above, with almost-instant response times and zero unscheduled downtime. (And the user story is posted on our website.)

    However, I wouldn't be so bold to say that the same thing is impossible with PHP .. just that it was relatively straight-forward to accomplish with Java/Tomcat/Coherence/MySQL.

    Again, I have to wonder why you would try to claim that PHP is better than Java or vice versa. They are two very different tools. I've rarely heard anyone compare Perl to Java, for example ..

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Shared Memories for J2EE Clusters
  114. Again, I have to wonder why you would try to claim that PHP is better than Java or vice versa. They are two very different tools. I've rarely heard anyone compare Perl to Java, for example ..Peace,Cameron Purdy
    PHP is not better than JAVA, it is just better for "typical" web applications. It is very popular to talk about "typical" or "80%" of use cases in JAVA, is not it ?
  115. the typical 80%[ Go to top ]

    IMO, the most common cases were the Java development goes wrong is not using the "stateless server" approach. If you need to cache session and state data to get acceptable performance you have over-complicated by a large factor already.

    The majority of typical web-applications, "80%", can be done using KISS stateless PHP with simple page-caching, letting the database take the load. For instance, I think that TSS could be done in simple PHP.

    It should be interesting to see were the "border/limit" really is in terms of transactions per second. Peter, care to give an estimate? Regarding connections I think it is very difficult for anyone to know exactly what is happening, cached (available) for a site under heavy load. That would require very deep knowledge of the database in question and processor registers, etc.

    Read the discussion at

    Friendster goes PHP

    "A typical Java application will make use of the fact that it is running under a JVM in which you can store session and state data very easily and you can effectively write a web application very much the same way you would write a desktop application. This is very convenient, but it doesn't scale."

    Regards
    Rolf Tollerud
  116. the typical 80%[ Go to top ]

    IMO, the most common cases were the Java development goes wrong is not using the "stateless server" approach. If you need to cache session and state data to get acceptable performance you have over-complicated by a large factor already.The majority of typical web-applications, "80%", can be done using KISS stateless PHP with simple page-caching, letting the database take the load. For instance, I think that TSS could be done in simple PHP. It should be interesting to see were the "border/limit" really is in terms of transactions per second. Peter, care to give an estimate? Regarding connections I think it is very difficult for anyone to know exactly what is happening, cached (available) for a site under heavy load. That would require very deep knowledge of the database in question and processor registers, etc.Read the discussion at Friendster goes PHP"A typical Java application will make use of the fact that it is running under a JVM in which you can store session and state data very easily and you can effectively write a web application very much the same way you would write a desktop application. This is very convenient, but it doesn't scale."RegardsRolf Tollerud

    Here is an old article I wrote back in 2002. The numbers are old and out of date, but the comparison between pool vs non-pool on page 15 shows the difference.

    http://jakarta.apache.org/tomcat/articles/performance.pdf

    the sudden drop in req/sec on graph 10 is because Oracle was running on a windows XP professional, which limits inbound connections to 10. On real server harder with Windows 2K3 Server, the numbers would be higher. On a single CPU, I'm going to guess it will plateau around 50 concurrent requests. On a 4 CPU box, with the same query it should plateau around 150 concurrent requests. The query is super simple though. It's just a select from a single table with like a couple thousand rows of data. The point of that test was to measure the cost of making new connections vs using an existing one.

    peter
  117. the typical 80%[ Go to top ]

    Read the discussion at Friendster goes PHP

    "A typical Java application will make use of the fact that it is running under a JVM in which you can store session and state data very easily and you can effectively write a web application very much the same way you would write a desktop application. This is very convenient, but it doesn't scale."

    First, I think that Friendster made the right decision. Their Java app sucked, and their PHP app works. What else matters? Is it possible that the complexity of Java contributed to their problems? Sure, I think it's possible. It's dangerous to ignore that possibility, or in the end Java will become as complicated as CORBA. ;-)

    Second, Java servers with user sessions and application state will scale just fine. In fact, for user sessions and read-only application state, it will easily scale linearly (using sticky load balancing), but the sessions won't have failover if a machine dies. Regarding read/write application state and user sessions with failover, we have customers that have scaled up to much (MUCH) larger loads than Friendster, and for applications that involve money changing hands (i.e. transactional as opposed to the FOAF-wannabe that you linked to.) In these cases, we've seen Java servers scale linearly from server number two well past server number 100, with each individual server handling more users than it could handle when the sessions were stored in a database, since accessing the database is also pretty expensive from the point of view of the server that is doing the access. Conclusion: Java is more scalable within a particular server, and Java is more scalable for adding servers, and even more so for stateful applications.

    $.02

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Shared Memories for J2EE Clusters
  118. Again, I have to wonder why you would try to claim that PHP is better than Java or vice versa. They are two very different tools. I've rarely heard anyone compare Perl to Java

    PHP is not better than JAVA, it is just better for "typical" web applications. It is very popular to talk about "typical" or "80%" of use cases in JAVA, is not it ?

    Yes, I agree with that, particularly for shared hosting environments (like the one we use).

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Shared Memories for J2EE Clusters
  119. Again, I have to wonder why you would try to claim that PHP is better than Java or vice versa. They are two very different tools. I've rarely heard anyone compare Perl to Java, for example ..Peace,Cameron Purdy
    PHP is not better than JAVA, it is just better for "typical" web applications. It is very popular to talk about "typical" or "80%" of use cases in JAVA, is not it ?
    I guess I've not done any typical web apps. Anyway, I try not to do web apps - but rather have applications with web UIs.
  120. Again, I have to wonder why you would try to claim that PHP is better than Java or vice versa. They are two very different tools. I've rarely heard anyone compare Perl to Java, for example ..Peace,Cameron Purdy
    PHP is not better than JAVA, it is just better for "typical" web applications. It is very popular to talk about "typical" or "80%" of use cases in JAVA, is not it ?
    I guess I've not done any typical web apps. Anyway, I try not to do web apps - but rather have applications with web UIs.
    "typical web application" is a registration form and almost static list of products and/or news/articles, probably it has some personalizied "welcome page" too. PHP is very good for this stuff.
    I am not a PHP fan, but hate to see "FUD" about non JAVA developers. Pool, cache, ORM or MVC existence for JAVA doe's not make JAVA developers smarter("server-side software engineering experts") than PHP developers or JSP better than PHP. PHP is a very pragmatic choise for many use cases too.
  121. Again, I have to wonder why you would try to claim that PHP is better than Java or vice versa. They are two very different tools. I've rarely heard anyone compare Perl to Java, for example ..Peace,Cameron Purdy
    PHP is not better than JAVA, it is just better for "typical" web applications. It is very popular to talk about "typical" or "80%" of use cases in JAVA, is not it ?
    I guess I've not done any typical web apps. Anyway, I try not to do web apps - but rather have applications with web UIs.
    "typical web application" is a registration form and almost static list of products and/or news/articles, probably it has some personalizied "welcome page" too. PHP is very good for this stuff. I am not a PHP fan, but hate to see "FUD" about non JAVA developers. Pool, cache, ORM or MVC existence for JAVA doe's not make JAVA developers smarter("server-side software engineering experts") than PHP developers or JSP better than PHP. PHP is a very pragmatic choise for many use cases too.
    Exactly the sort of "web app" that I would use it for - if I did those. I am not a graphic artist so I don't do those. At least I have not done any yet.
  122. I wouldn't say that performance is the **major** problem with PHP.

    At first, it's maintainability that hurts. Some time back a friend of mine asked me to help making little change to one of PHPBB clones (for those non-PHP types: this is popular online forum software). The change was, namely, authentication, which has to come via another table so that forum could share user base with another app. My, my, that was tough! It's about 20+ places scattered all around these .php files. So I've spent about 5 hours changing it and then fixing it after it broke here and there again.
    I wouldn't say that PHP is that bad (one could have written the same thing in JSP), but rather concept of 'everything per request' promotes such lame code practices, resulting in unmaintainable code.

    The other big problem of PHP is security (again, not PHP but coding practices that it encourages). Many PHP packages (having version 7.x and more!) are made in such a way, that they have to write something to some directory under site root. In other words - the user web server process runs under must be allowed to write files to filesystem. That results in numerous security exploits simple enough for teenagers to play with.
  123. Probably most PHP applications do not use persistent connections (connect once per request), it is more scalable than pools with backround threads, fake query execution,reconnection and abandon stuff. LAMP doe's not has problems with pools and caches, because smart PHP developers know how to develop scalable sofware without workarounds. I respect "JAVA camp" arogance and ignorance, but it is a good idea to learn competitor tools too.

    This post shows a real lack of understanding of server-side software engineering. I hate to be this rude about it but what you have written here is foolish and naive. It clearly demonstrates that you do not understand the concept of scalability. Create a connection per request is a terrible way to do things and will make it impossible to scale the application. DB's will only support so many connections.

    Furthermore, there's nothing in Java that prevents one from creating a Connection per request. Experienced and competent developers use pools and caches because of the benefits they provide.
  124. why should it matter[ Go to top ]

    Probably most PHP applications do not use persistent connections (connect once per request), it is more scalable than pools with backround threads, fake query execution,reconnection and abandon stuff. LAMP doe's not has problems with pools and caches, because smart PHP developers know how to develop scalable sofware without workarounds. I respect "JAVA camp" arogance and ignorance, but it is a good idea to learn competitor tools too.
    This post shows a real lack of understanding of server-side software engineering. I hate to be this rude about it but what you have written here is foolish and naive. It clearly demonstrates that you do not understand the concept of scalability. Create a connection per request is a terrible way to do things and will make it impossible to scale the application. DB's will only support so many connections.Furthermore, there's nothing in Java that prevents one from creating a Connection per request. Experienced and competent developers use pools and caches because of the benefits they provide.

    why should making a new connection to the database matter? It's not like it takes at minimum 50ms to create a connection when the server is idle and longer when it's under load. It's not like all the jdbc drivers out there all support connection pooling. why should connection pooling matter? Surely microsoft implemented connection pooling for no good reason.

    For those humor impaired, I'm just joking. anyone can easily see the effect of not using db connection pooling by sending 100 concurrent requests to a webserver in both modes. In the non-pooling mode, the bottleneck is reached very quickly and results in db connection errors. Even the simple benchmarks I ran with tomcat + oracle back in 2002 easily show that.

    though i suppose in some alternate reality, it might be possible to create a new connection for each request and have it scale equally well as connection pooling. until then, webservers are stuck in this reality and have to obey the laws of physics.

    peter
  125. why should it matter[ Go to top ]

    anyone can easily see the effect of not using db connection pooling by sending 100 concurrent requests to a webserver in both modes. In the non-pooling mode, the bottleneck is reached very quickly and results in db connection errors. Even the simple benchmarks I ran with tomcat + oracle back in 2002 easily show that.though i suppose in some alternate reality, it might be possible to create a new connection for each request and have it scale equally well as connection pooling. until then, webservers are stuck in this reality and have to obey the laws of physics.peter
    Yes, the best way is just to test stuff and to use the most scalable configuration (it scales better without pool in some of cases).
    BTW I asume you dissable DNS resolving or map stuff in "hosts" file before to test.
  126. where exactly lies the bottleneck?[ Go to top ]

    Why do we not do our own test?. Peter works in another realm of reality were 2000 trans/per second is not heavy load and is counted as a "small" site. So when he says "the bottleneck is reached very quickly" by that he doesn't mean what most people mean by that.

    I will do my own tests. But I need another piece of information. How common is websites with less than 2000 trans/per second (or 30 simultaneous TSS forums)?

    I'll be back
    Rolf Tollerud
  127. Connection pools solve many problems, but it can produce many problems too. I prefer to configutre and to tune sofware before to blame or to use workaround. Try to tune network parameters, DNS stuff, probably some vendor provided solution can help too http://www.stanford.edu/dept/itss/docs/oracle/9i/network.920/a96580/mts.htm
  128. OK, thanks for the link.
  129. If you want to outperform JAVA workarounds then try some dirty way for this test too http://httpd.apache.org/docs-2.0/mod/mod_cache.html
  130. where exactly lies the bottleneck?[ Go to top ]

    If you want to outperform JAVA workarounds then try some dirty way for this test too http://httpd.apache.org/docs-2.0/mod/mod_cache.html

    one can easily make PHP scale just as well as Java, by generating static HTML for the most heavily viewed pages. mp3.com used that approach back in 2000. They setup each front end server with mysql and apache. On a regular cycle, the backend would generate new pages and push them out to the front end servers. The front-end databases kept user data, but the non-user specific pages were static files (ie music, categories, etc). They simply wrote their webapp to lookup the static file. When Oracle 8i was released, it came with a tool to create batch jobs that generate HTML. I don't know if people used it or how popular it is, but the concept is scalable. It's not appropriate for real-time stuff, but then most apps don't need or care about real-time. this technique can be used with Perl, PHP, Java, or ASPX, so it's not a platform thing. It's a design choice/decision.

    for something like MP3.com, it would have been silly to make it real-time and run all those queries against the database when generating static files is more than sufficient.

    peter
  131. where exactly lies the bottleneck?[ Go to top ]

    Peter,

    "batch jobs that generate HTML"

    Yes that is an interesting teqnique that is underused I think. Also the nice thing about it is that you can use XML-XSLT parsing to your hearts content! But back to the connection pool thing.

    What is meant by the term "connection pooling" according to you? Back in the old days when I was working with Java we used a library that was called Poolman. That is what I mean by connection pooling, a Java library that keep a graph of connections object in memory.

    But nowadays I see connection pooling in the JDBC driver, as a functionality in Oracle Shared Server architecture, build into SQL Server, as wrapper JDBC Driver classes, PHP has persistent connections, etc. What is it all? Can everything be used together? Are Java libraries like Poolman still used?

    I see no area that is so completely littered and confused. I see at more than a dozen different types of connection pooling. Please clarify. What do you use in the simple normal typical mainstream web application?

    And please. No XML-parser, no financial-aggregations, no special anything, just in a normal typical mainstream web application with medium load. Try to keep to the subject for once, if you can! :)

    Regards
    Rolf Tollerud
  132. Peter,"batch jobs that generate HTML"Yes that is an interesting teqnique that is underused I think. Also the nice thing about it is that you can use XML-XSLT parsing to your hearts content! But back to the connection pool thing.What is meant by the term "connection pooling" according to you? Back in the old days when I was working with Java we used a library that was called Poolman. That is what I mean by connection pooling, a Java library that keep a graph of connections object in memory.But nowadays I see connection pooling in the JDBC driver, as a functionality in Oracle Shared Server architecture, build into SQL Server, as wrapper JDBC Driver classes, PHP has persistent connections, etc. What is it all? Can everything be used together? Are Java libraries like Poolman still used?I see no area that is so completely littered and confused. I see at more than a dozen different types of connection pooling. Please clarify. What do you use in the simple normal typical mainstream web application?And please. No XML-parser, no financial-aggregations, no special anything, just in a normal typical mainstream web application with medium load. Try to keep to the subject for once, if you can! :)RegardsRolf Tollerud

    for what it's worth, here my quick definition of connection pooling.

    1. a pool of database connections, which can be reused
    2. the connection pool is actively managed
    3. preferrably implement the JDBC pool API

    Back on the topic of PHP, if Zend implements connection pooling, I see no technical reason why PHP can't get closer or match Resin + JDBC + connection pool. A solid connection pooling driver would go a long way and make PHP more robust. If I had time, I'd write one myself, but I have no time these days. it's getting to the point where, the tools you choose depends on your skill sets and performance in many cases makes absolutely no difference. It's only when you get to the edge cases that you start needing other features. but I see no reason to polute PHP with that stuff.

    peter
  133. ?[ Go to top ]

    Your answer is so general that it helps nothing. It is like I've asked "what is a Orchidé" and you say: "It is a flower". I am no wiser than before. That I was interested in was to sort out the differences between the different types of connection pooling.

    For instance you say "but assuming the jdbc driver is relatively bug free, a simple connection pool will generally be faster than making new connections."

    So you imply that the current jdbc connection pool is sufficient. Why then do Jakarta has the DBCP (Database Connection Pool Configurations)? From their site,

    "DBCP provides support for JDBC 2.0. On systems using a 1.4 JVM DBCP will support JDBC 3.0". Why bother if there is build in connection pooling in JDBC 2.0/3.0?

    On a comical note: What happens if you have the JDBC pooling turn on and still use DBCP and perhaps Oracle Shared Server Connection Pooling? :)

    Regards
    Rolf Tollerud
  134. never used Jakarta DBCP[ Go to top ]

    I'm totally bias here, but I've never used DBCP and prefer to use the vendor provided database connection pooling. I'm not sure what detail you're looking for. Of the jdbc drivers I know of, and have looked at in some detail, they all implement connection pooling in similar ways. Back in 2000/2001 the differences between jdbc drivers varied more, but by now they are all pretty similar. There's only a few ways to write an efficient connection pooling jdbc driver.

    I honestly couldn't tell you what would happen in the case you described. Generally speaking, you don't really turn on connection pooling. Tomcat's server.xml doesn't give you a way to configure a pool. If you're using weblogic, it's configured in config.xml. In the case of weblogic, you define jdbc connection pools with <JDBConnectionPool CapacityIncrement="5" .... /> Weblogic will set the pool parameters and initialize the pool.

    http://cvs.apache.org/viewcvs.cgi/jakarta-commons/dbcp/doc/ManualPoolingDriverExample.java?view=markup

    with DBCP, you have to configure the webapp, but the underlying driver "should" still be the vendor provided driver. does that answer your question?

    peter
  135. duh on my part[ Go to top ]

    after I hit reply, I remember you can configure tomcat to create a pool.
    http://jakarta.apache.org/tomcat/tomcat-5.0-doc/jndi-datasource-examples-howto.html

    it still uses the vendor provided driver, so DBCP is just a wrapper. I really should proof read and double check facts before I hit reply. oh well, bad habits are hard to break.

    peter
  136. ?[ Go to top ]

    Your answer is so general that it helps nothing. It is like I've asked "what is a Orchidé" and you say: "It is a flower". I am no wiser than before.
    The porridge is too hot. The porridge is too cold. When is the porridge gonna be just right for you?
  137. ?[ Go to top ]

    Your answer is so general that it helps nothing. It is like I've asked "what is a Orchidé" and you say: "It is a flower". I am no wiser than before.
    The porridge is too hot. The porridge is too cold. When is the porridge gonna be just right for you?

    I don't find any fault in Rolf's question. He probably has never written a connection pooling driver, so it may seem like black magic. For those who have written connection pooling drivers, it's obvious. Before I had to write a pooling driver, it was a bit of mystery to me too. It's a valid question. It's easy to forget that others may not have the experience and can't tell the difference.

    peter
  138. ?[ Go to top ]

    Peter, I have to disagree with you. One shouldn't need to write a connection pool just in order to understand how it works or what's the concept behind it.

    PS: I think it is such a simple concept I am leaning towards Rolf is just playing one of his troll games with you, but anyway, have fun having a technical talk with him!

    Regards,
    Henrique Steckelberg
  139. Actually I have written a connection pool class once, and in Java too. But I have to admit that it merits was questionable! :)

    What puzzled me was "JDBC supports connection pooling". Did it meant that you could just go on and use JDBC and still get connection pooling automatically, just by setting a parameter? Of course not.

    So after surfing an hour I can assure you that the majority of Java developers don't know either. Extreme confusion reigns.

    JDBC are not enough you still have to have your connection broker. All app containers have their own connection broker. In tomcat it is called DBCP. But being Open Source, you can use DBCP outside Tomcat.

    So what it boils down too is that if for example you use Weblogic with Oracle, then you have several choices,

    1) You can use Weblogic Connection pooling
    2) You can use a local DataSource with DBCP, (or other third part product), but in that case you loose the ability to distributed transactions that span more than one database.
    3) You can use Oracle Connection pooling

    If you use a local DataSource you don't have to care about the specific quirks and anomalies of the target container's connection pool -- you have your trusted pool that you've developed and tested on.

    With Spring you can expose a DataSource dependency in your components and pass in a specific DataSource instance at initialization time. That way, switching from a local DataSource to a JNDI one or vice versa is just a matter of configuration.

    No 2) already exist with with PHP and is called persistent connections. You also can use 3).

    A charlatan makes obscure what is clear; a thinker makes clear what is obscure.
    —Hugh Kingsmill

    Regards
    Rolf Tollerud
  140. P.S. I forgot[ Go to top ]

    Thanks to Juergen Hoeller.
  141. Henrique is the most smart of you[ Go to top ]

    ...But I have to admit that it merits was questionable!
    I'd put money on it! Anyone want in? :P
    A charlatan makes obscure what is clear; a thinker makes clear what is obscure
    Finally. A quote that actually makes sense and sort of goes with the topic.

    BTW, Dictionary.com says a charlatan is "A person who makes elaborate, fraudulent, and often voluble (Marked by a ready flow of speech; fluent) claims to skill or knowledge; a quack or fraud." Sounds vaguely familiar. Hmmm.
  142. Henrique is the most smart of you[ Go to top ]

    Actually I have written a connection pool class once, and in Java too. But I have to admit that it merits was questionable! :)What puzzled me was "JDBC supports connection pooling". Did it meant that you could just go on and use JDBC and still get connection pooling automatically, just by setting a parameter? Of course not.So after surfing an hour I can assure you that the majority of Java developers don't know either.

    I don't know what java developers you've worked with, the ones I know first hand know what connection pooling is and have experience using it. everyone experience is different obviously.
    Extreme confusion reigns. JDBC are not enough you still have to have your connection broker.

    I'm not sure what you're calling connection broker here. It's really just an initialization process. weblogic does not actively manage the connection, unless you're talking about weblogic T3. if by broker you mean something that reads the configs and sets the parameters in the underlying driver provided by the manufacturer, then I would say yes. the driver is just a driver, something has to initialize it.
    All app containers have their own connection broker. In tomcat it is called DBCP.

    DBCP isn't necessarily the official pooling driver/manager. Atleast I've never considered it the official one.
    But being Open Source, you can use DBCP outside Tomcat.So what it boils down too is that if for example you use Weblogic with Oracle, then you have several choices,1) You can use Weblogic Connection pooling 2) You can use a local DataSource with DBCP, (or other third part product), but in that case you loose the ability to distributed transactions that span more than one database.3) You can use Oracle Connection pooling
    If you use a local DataSource you don't have to care about the specific quirks and anomalies of the target container's connection pool -- you have your trusted pool that you've developed and tested on. With Spring you can expose a DataSource dependency in your components and pass in a specific DataSource instance at initialization time. That way, switching from a local DataSource to a JNDI one or vice versa is just a matter of configuration. No 2) already exist with with PHP and is called persistent connections. You also can use 3).A charlatan makes obscure what is clear; a thinker makes clear what is obscure.—Hugh KingsmillRegardsRolf Tollerud

    I wasn't aware PHP already had connection pooling that is functionally equivalent to JDBC connection pooling and ADO.Net connection pooling. thanks for the info.

    given that PHP has connection pooling, than I see no reason why it can't scale as concurrent load increases. Assuming you haven't reached the limits of your database. Once you've reach the limits of your database in terms of concurrent queries/tranactions, you have to use alternate approaches to scaling. In those cases, I stand by my assertion that the same techniques used on J2EE or .NET are also applicable to PHP. It's just a matter of whether a person decides it's worth while to write their own infrastructure to do it.

    peter
  143. more like Debussy than Bach[ Go to top ]

    Peter,

    I don't doubt for a second that you are a very knowledgeable fellow. Extreme so in fact.
    But you just do not have the ability to explain "so even 7 years old can understand". Sorry! :)

    Regards
    Rolf Tollerud
  144. don't remind me[ Go to top ]

    as much as I try, I end up in geek speak. you're not the first to tell me that and I doubt you'll be the last. that's ok, I have the rest of my life to learn how to speak like normal human being. I seriously doubt I'll make any progress though. All those years in college has warped my delicate brain.
  145. don't remind me[ Go to top ]

    Well, he did it to you again. It is tough, if not impossible to explain something purely technical in layman's terms.

    Not sure why he is complaining. He does the same except with Literature.
  146. don't remind me[ Go to top ]

    "Well, he did it to you again. It is tough, if not impossible to explain something purely technical in layman's terms."

    No, it is not. Just read something by Juergen Hoeller.

    It is you that is complaining, Mark. But you are only jealous because you're not insulted by Hani!

    Regards
    Rolf Tollerud
  147. don't remind me[ Go to top ]

    "Well, he did it to you again. It is tough, if not impossible to explain something purely technical in layman's terms."No, it is not. Just read something by Juergen Hoeller.It is you that is complaining, Mark. But you are only jealous because you're not insulted by Hani!RegardsRolf Tollerud
    If it was that easy, everybody would speak as clear as Juergen. Very few do, which is a sad thing.
  148. don't remind me[ Go to top ]

    "Well, he did it to you again. It is tough, if not impossible to explain something purely technical in layman's terms."No, it is not. Just read something by Juergen Hoeller.It is you that is complaining, Mark. But you are only jealous because you're not insulted by Hani!RegardsRolf Tollerud

    I am not complaining (one would think you, of all people here, would understand the term).

    But am jealous. :)

    Juergen Hoeller may explain things well, but I don't believe the technologically challenged(layman) will understand. Most of those here are not laymen.
  149. PHP persistent connections[ Go to top ]

    "Well, he did it to you again. It is tough, if not impossible to explain something purely technical in layman's terms."No, it is not. Just read something by Juergen Hoeller.It is you that is complaining, Mark. But you are only jealous because you're not insulted by Hani!RegardsRolf Tollerud

    I decided to take a closer look at persistent connections in PHP and found this page http://terra.di.fct.unl.pt/docs/php/features.persistent-connections.php.htm.

    If I understand it correctly, PHP persistent connection is closer to an object pool. In other words, it's just some connections in a pool and the PHP engine does not actively manage them. I don't know if that page is correct or not, but according to the description, persistent connections are just open connections.

    That is quite different from JDBC connection pooling, which is actively managed. PHP connection pooling is good for reads, since it lacks proper support handling auto commit. If I were to do perform 2 requests to a page that uses persistent connections and forget to call commit, there's a chance those transactions will be lost. If a third request comes in and does a rollback, I would loose all 3 inserts. If the database connection dies, the records would be lost.

    it puts more burden on the developer to check everything works correctly and have a good set of functional tests to make sure a new release handle transactions correctly. in skilled hands, PHP persistent connections can be used in the same way as stock JDBC connection pooling. The difference is an important consideration in my mind. just some random thoughts on the differences.

    peter
  150. PHP persistent connections[ Go to top ]

    Since it seems people refuse to acknowledge that comparing PHP to Java is like comparing apples and bannanas (yeah, oranges, at least have similiar shape), here is the thing:

    Disregarding PHP shortcomings, that one can argue, what is the major difference between the two as far as web-development goes (that is - the only place you can compare them)?

    OK, I am gonna oversimplify here, so do not get anxious on me, but here is what it boils down to:
    1) PHP, by far, is a parser that lives in a web-server (e.g. Apache) process. Lifespan of a php process is limited to that of an HTTP Request. PHP does not have its own threads, it does not have its application scope.

    2) On the contrary, Java/J2EE applications are full-blown, quite independent, applications for which HTTP Request is just an event that it responds to.

    Because of this, even if you compare JSP to PHP, it is still apples and oranges as JSP is HTML-embeded/scripting way of doing things but still #2, above, while PHP is unable to leave the constraints of #1.

    This is the fundamental difference. The rest is just consequences that are fixable and comparativly minor.

    Using a parser (even if highly sophisticated and wanna-be object-oriented) may seem easier than writing an application, but - for who?

    You know, I have a deep belief that, by far, the only reason why PHP became so widespread is that there are so many cheap, shared hostings available on PHP. That's the real reason, not anything related to its internal beauties. People had a simple choice - for that price, either have a statis HTML home-page, or a dynamic PHP website that can handle only that much but still - can.

    And yes - making PHP HTML-embeded made it much more attractive, for the beginners, than the then-cumbersome Perl.

    Try google-ing for cheap PHP hosting and then try Java or .NET hosting. You will have your answer on - why PHP is so popular.
  151. Thumbs up for PHP[ Go to top ]

    I have been a fan of PHP for quite some time now, PHP has such good potential to beat Java in web development.
  152. Henrique is the most smart of you[ Go to top ]

    In those cases, I stand by my assertion that the same techniques used on J2EE or .NET are also applicable to PHP. It's just a matter of whether a person decides it's worth while to write their own infrastructure to do it.peter
    It is possible to use the same "J2EE techniques" with PHP too (OOP,ORM,MVC,pools,caches) and this stuff is used, it is possible to scale JAVA applications using "per request" architectures too (optimization/workaround must "transparent" anyway).
    You can prove it yourself using google.
  153. "SQL Relay is a persistent database connection pooling, proxying and load balancing system for Unix and Linux."

    http://sqlrelay.sourceforge.net/

    Regards
    Rolf Tollerud
  154. ?[ Go to top ]

    Your answer is so general that it helps nothing. It is like I've asked "what is a Orchidé" and you say: "It is a flower". I am no wiser than before.
    The porridge is too hot. The porridge is too cold. When is the porridge gonna be just right for you?
    I don't find any fault in Rolf's question. He probably has never written a connection pooling driver, so it may seem like black magic. For those who have written connection pooling drivers, it's obvious. Before I had to write a pooling driver, it was a bit of mystery to me too. It's a valid question. It's easy to forget that others may not have the experience and can't tell the difference.peter

    Peter. Its not the questions, so much as the comments. He says you are being to detailed, then says you are being way too simplistic. I don't think you are going to "win" with Rolf. As soon as you do catch him on anything he will 1: ignore you or 2: Come up with some big quote that makes sense only to him and leave you scratching your head with no way to respond. Kind of like the joke - "How many pancakes does it take to shingle a doghouse?" "Six, because the ice cream doesn't have bones."
  155. where exactly lies the bottleneck?[ Go to top ]

    Connection pools solve many problems, but it can produce many problems too. I prefer to configutre and to tune sofware before to blame or to use workaround. Try to tune network parameters, DNS stuff, probably some vendor provided solution can help too http://www.stanford.edu/dept/itss/docs/oracle/9i/network.920/a96580/mts.htm

    absolutely agree on that. When oralce 8i first came out, their jdbc drivers didn't support connection pooling and there was a bug in the driver where connections would become unusable after X queries. We ended up writing our own connection pool, which would periodically recycle the connections, until Oracle released new drivers fixing that connection bug.

    but assuming the jdbc driver is relatively bug free, a simple connection pool will generally be faster than making new connections. On more than one occasion, users have blamed tomcat for memory leaks when it was something simple like forgetting to call connection.close(). It's easy to make a mistake and kill your database, but that is a development issue in my mind and not a technology issue.

    peter
  156. This post shows a real lack of understanding of server-side software engineering.
    I am very happy to discuss with server-side software engineering expert, thanks.
  157. Wanna bet?[ Go to top ]

    $100 bucks says Rolf is the "un-named analyst."

    Sorry, that was way too /.
  158. Comparing Apples with Pears[ Go to top ]

    One cannot compare an Enterprise Framework with a scripting language. It's like comparing apples with pears.

    As already mentioned, the only true comparison is PHP v JSP, and from what I have evidenced, JSP wins hands down. I've recently integrated WordPress (blog) and Coppermine (photo gallery) into my web site. Both written in PHP. Coppermine was a nightmare.. took me two days of hacking to integrate it into my web site template and I still can't produce valid XHTML. WordPress is great, but if something goes wrong, I get a SQL trace on my screen. No true error handling.

    To me PHP is a step back to my days as an ASP developer... just run time scripting. You intersperse HTML with PHP code/ calls to PHP functions.

    But could you render an HTML table from a database as neatly as a taglib like displaytag does? No.

    JSP can be written like PHP, or it can be taken to the next level - MVC, and then, if required, it can front an n-tier solution. PHP is single-tier, non-MVC and parsed at run time.

    But why is PHP so popular?

    1. Because it's easy to learn. Just like ASP is. Perhaps JSP is not/ was not marketed well enough? Same with ColdFusion - great scripting language- years ahead of JSP taglibs.

    2. Because it's easy for hosts to maintain. Though setting up Resin, and with a bit more effort, TomCat is a breeze, managing the container and the JVM for virtual hosting etc. requires the hosting company to have Java (expensive) admin skills.

    3. There is a momentum behind the language. Lots of good (as in functionality) software is being produced - blogs, bulletin boards etc. It's easy to hack. No complex MVC config to get one's head round.

    The problem with Java is that once you've moved up from scriptlets to taglibs, MVC and SOA, you never turn back, so when your friend wants a simple web solution to post last weeks amature footie results to their site, you deliver a full-blown solution, rather than throwing together a JSP page with an Apache SQL taglib.

    Personally, I think PHP has a long way to go. Many PHP developers seem to be impressed by ruby on rails. And rightly so. It encompasses the best practices of Java web development, but offers nothing new.

    Why are these people impressed by Ruby, but never gave Java a chance? Marketing I guess. Perhaps what's needed are some tutorials and books teaching web scripters how to write web sites just using taglibs.

    [ apologies for the essay! first post. too enthusiastic ;-) ]
  159. PHP has a strong and active community so makes sense to work with them.

    IMHO, core domain objects (POJOS) should be write in Java by professionals with string OO / domain modeling skills. Domains models should be well tested and documented, in a very professional way. And any script language should be able to leverage theses domain models to produce a real application.

    JSR 223 can help with this. These dynamic languages should be used to write application specific (container dependent) code like the view and the controller layers.
    "People recognize that scripting languages fill a certain space in the market, including others in the Java community," Murphy said. "As IBM pushes into the small and medium-size market, which uses a lot of scripting, they need to make a play there."

    "IBM has made significant investments in Java software and will continue to invest in Java industry standards and its WebSphere-branded line of server software and tools, Smith said. Its commitment to PHP is a way to reach out to more developers, particularly in small and medium-size businesses, he said"

    "Smith said that the simplicity of PHP is one of its greatest assets. Business analysts without a background in computer science, for example, could create Web pages that pull data from back-end sources, he said"

    Seems like they are obsessed about things like simplicity, reach more developers an so on. For me, it make sense: these application developers without strong OO skills should have a chance to work within “enterprise applications” but also they should at least be able to leverage the domains written in Java. Actually, sometimes a business user should be able to write some scripts...

    We already have JVM based script solutions like http://www.beanshell.org, http://pnuts.dev.java.net and off course, http://groovy.codehaus.org, I hope any of it gain more momentum following this IBM move.

    Interesting, Pnuts was developed by Sun Japan...

    BTW, I found Pnuts on this blog entry
  160. JSR 223 can help with this.

    Thank you for mentioning this!: I'm surprised that this didn't come up in the discussion at the start (but then, this is not really a discussion, is it ;-) given that JSR 223 recently published its public review draft and explicitly aims to support PHP as a scripting language on the Java platform. Suddenly, the world is not so easy to divide into Java and non-Java, is it...and IBM's move makes total sense - business sense and technical sense.

      cheers,
      gerald
  161. I guess by "for years" the author means the years IBM was supporting Java in its infant stages.

    oh please! Looking for languages they could buy without fighting Microsoft, sure. They dont fear SUN so much.

    As if IBM would be investing so hard in Linux if it werent for Java and their best baby eclipse.

    Anyone remember the C++ wars over the APIs and IDE's? Guess not. To be continued....

    Next story IBM supports ANSI C and ANSI COBOL. For the better of everyone.
  162. PHP is more easy ¿REALLY?[ Go to top ]

    ¿What is the difference between a Web application with Tomcat- JSP-JDBC with NO classes and a PHP version?

     Essentially NONE.

     Garbage code can be done easily with Java too, in fact, JDBC is by far more easy and portable than the nightmare PHP database API.
  163. PHP is more easy ¿REALLY?[ Go to top ]

    They are not same, and Java is better.
    Why?

    Because pages written PHP is NOT portable, not cross platform , and not cross server.

    Many web server outthere are support PHP, but they are not same. Every server are difference, even one little configuration/version mis-match may cause your PHP code fail.

    For example,
    some server reduce numbers of modules for PHP functions.
    do some configuration in php.ini.(e.g. global=on/off)

    People have to confirm the env. info. provided by phpinfo() to make sure it works.

    And, you cannot ask you web hosting provider recompile the server, and let the required modules for you.

    In this case, Java is easier.
    1st, the version of JRE and Servlet/JSP version are standard.
    2nd, you can easily add you jar file youself if needed.
         (not include hundred jsp/php file)
    3re. JSP provided more.
         e.g. taglib, EL(I hate tag[server] in tag[html]),
         codebehide
    4th. Servlet provided more.
         e.g. filters, authorization..etc
    5th. Non-JSP Java web solution (e.g. tapestry, W4T)
    6th. Java is not only the Web.
         many techology are supported,

    Java can do everything PHP can.
    But, PHP cannot do manythings JSP/Servlet/J2EE/Java did.

    Script is Script.
    If wanted to do simple things in WEBPAGE, every techology can, ofcause people can pick a script language.
    But, if people wanted to build a website application, i don't think PHP is a good choice.
  164. PHP is more easy ¿REALLY?[ Go to top ]

    php is fast to code.
    php does not need to be compile - and This is a huge loss of time when you create a complex html page (But in fact now that I use css, this simplifies a lot of things)
    php can be found everywhere
    you can create a front end with the best forum, the best wiki, the best cms, the best blog machine in less than 20 minutes.
    The communauty is huge, the library is huge, the products are well solid.

    I tuely agree, Jsp IS better but has never proved it yet.
  165. PHP is more easy ¿REALLY?[ Go to top ]

    php is fast to code.php does not need to be compile - and This is a huge loss of time when you create a complex html page

    I have found that not being able to compile is a great loss of time. I've spent more time trying to find a scripting error than compiling. And if you're page is that complex, it should be broken down into components. That is part of the page paradigm problem.

    I've done plenty of scripting - Javascript, XSL, ASP, PHP, some mainframe "scripting", etc. - and I found that anything beyond the very simple becomes very complex, very quickly. If one likes not have strongly typed, no checked exceptions and no compiling then fine. But it definitely is not better.
  166. PHP is more easy ¿REALLY?[ Go to top ]

    php is fast to code.php does not need to be compile - and This is a huge loss of time when you create a complex html page (But in fact now that I use css, this simplifies a lot of things)php can be found everywhereyou can create a front end with the best forum, the best wiki, the best cms, the best blog machine in less than 20 minutes.The communauty is huge, the library is huge, the products are well solid.I tuely agree, Jsp IS better but has never proved it yet.

    Intresting you use CSS to simplify a lot of things, means ???. I think it should be "Use CSS to reduce HTML"
  167. PHP is more easy ¿REALLY?[ Go to top ]

    Your right, I meant "reduces html". But then I can spend less time on the html pages.
    The main idea is not to deal with presentation with html tags but to have the lightest, simplest html, which means simpler jsp (or velocity, or struts, or jstl, or... php). Just keeping the sense of tags, the meaningfull tags. If there is a list of links, just add >ul< >li< ; a table of results : just one table etc... half a dozen of "ids" and "classes" and that's it. No more table imbrication, or any complex html problems. And when a correct html page is produced by my jsp, the css will take control of the design. Sometime i say -just to make fun- it is inversion of control applied to design.
  168. Having used Java for over five years I have to say it is still far from perfect on the web tier (JSPs, Struts, Tapestry... none are ideal). However, I believe it is unrivalled on the server tier, especially with OO-Relational mapping tools like Hibernate and the various caching technologies available to plug into it. It is simple to write server-side Java code with Hibernate and web services (Java should not be thought of as *just* EJBs). Deploying to a JBoss server, for example, is much easier than one might think.

    On the web tier, PHP with a templating technology like Smarty is very fast to develop with and very flexible when dealing with the "mushiness" of web clients. With the simple SOAP support in PHP 5 one can deploy a PHP web tier on Apache and then communicate over SOAP with a fast, lightweight, transactional Java server architecture.

    Integration is the key, no one technology will be best of breed in all areas.

    Just some thoughts on the subject,

    Sean.
  169. Hi Guys ,

    PHP has it's place with the MS development crowd . Just as Java was the next evolution for the C++/smalltalk coders , PHP is meant for the ASP VB developers .

    Think of it like this : would my local pizza shop have a java programmer create it's web site to display specials and offer email alerts or would they use ASP / PHP / perl to create the web site .


    along the same lines : would my bank use PHP to help me mannage my account or would they use a heavy hitter like java/j2ee ?
  170. You know what this is about - IBM is setting a threating tone aimed directly at Sun.

    The bottom line is - We are opening up tons of our software that is going to live on Sourceforge.com. We are giving away the patents. YOU DO THE SAME with JAVA.

    If Sun is not willing to do that, then looks like IBM may shift its weight towards other areas like PHP or, god knows what software, utlimately spelling a slow demise of JAVA.

    Unfortunately, at the end of the day, money loosing Sun is no match for IBM. IMHO Sun should start working with IBM rather closely and start doing things like - Get those IBM projects hosted on Java.net, get rid of Netbeans suite and put those $$$ in eclipse, open up Java in mutally agreeable fashion with inputs from development community, and ...

    So, I guess when this happens, IBM doesnt have to run around like a goose with its head cut off, making announcements like we are going to support PHP or, god know what software
  171. If Sun is not willing to do that, then looks like IBM may shift its weight towards other areas like PHP or, god knows what software, utlimately spelling a slow demise of JAVA.

    I don't think you realise how much IBM has invested in Java, and how much intensive research IBM has put into Java technology and VMs. This is a minor debate over web front ends, and has no impact on the future of Java in IBM or anywhere else.
    get rid of Netbeans suite and put those $$$ in eclipse

    Absolutely not. Why do we want a single IDE to dominate? Competition is good. NetBeans has a strong and growing user community, as do other products.
    So, I guess when this happens, IBM doesnt have to run around like a goose with its head cut off, making announcements like we are going to support PHP or, god know what software

    IBM has been doing this for years. This statement about PHP has, I believe, a lot less significance that many are giving it.
  172. PHP is security nightmare[ Go to top ]

    Just look at how many security problems with PHP (and software packages that're built on it, eg. phpbb) in the last few years ... and upgrading to newer version is not easy esp. if one (like me) has a lot of boxes to upgrade.

    PHP is fine to be a scripting language for the web, but I'm beginning to lose faith on it because of all these security problems. Same can be said to Apache httpd (and openssl for that matter).
  173. Java is great, JSP sucks[ Go to top ]

    I love developing in Java. Unfortunately, in Java, doing complex things is not so hard but doing simple things can be incredibly hard. In my mind the (lack of) quality of error reporting, bad documentation and the fact that JSP just sucks play an important role.

    The quality of JSP and particularly error reporting and IDE support dor JSP are the biggest showstoppers. And there is a simple solution: look at Cold Fusion and do as they did.

    Developing in CF is a breeze. Many people don't like it and argue that it invites bad coding (SQL in script page etc.) and that's true. However, for people who know what they're doing this is not an issue. In my mind, JSP is the backward little brother of CF. With CF, there is a well documented API, solid IDE support so you don't need to jump from this to that site to get the tag descriptions, and the error reporting is actually helpful.

    When I look at the many hours I have struggled doing stupid simple things in JSP, getting "there is an error"- type error messages, not to mention the very informative: "unable to compile JSP", and working with half-baked custom tag libraries with varying support for expressions and no support at all for getting things done, I can well understand a desire for something effective like PHP or Cold Fusion.

    It's just: why not put the effort in JSP and get it right?

    Cheers,

    Marc
  174. Java is great, JSP sucks[ Go to top ]

    The quality of JSP and particularly error reporting and IDE support dor JSP are the biggest showstoppers.

    How many IDE's supporting PHP are there out there, really?
    And there is a simple solution: look at Cold Fusion and do as they did.Developing in CF is a breeze. Many people don't like it and argue that it invites bad coding (SQL in script page etc.) and that's true. However, for people who know what they're doing this is not an issue.

    If you absolutely want to, you can have as much sql in your jsp-scripts as you like. But, if you know what you're doing you *dont* want that. Period. For getting small stuff up and running very fast mixing data and presentation layer works very well, though. In the keeping-warm-by-wetting-ones-pants way ...

    In my mind, JSP is the backward little brother of CF. With CF, there is a well documented API, solid IDE support so you don't need to jump from this to that site to get the tag descriptions, and the error reporting is actually helpful.

    In a way you're right. CF is running on top of J2EE these day's, you know. But, jsp is well documented, and there is solid ide support out there. What the situation is like for PHP I'm not aware of.
    When I look at the many hours I have struggled doing stupid simple things in JSP, getting "there is an error"- type error messages, not to mention the very informative: "unable to compile JSP", and working with half-baked custom tag libraries with varying support for expressions and no support at all for getting things done, I can well understand a desire for something effective like PHP or Cold Fusion.It's just: why not put the effort in JSP and get it right?Cheers,Marc

    First of all, getting errors at compile time is a *good thing*. But I agree, when the jsp-page is turned inside-out before it is compiled (as a servlet) it's kinda confusing trying to understand compiler error messages. However, there are tools that will help you with this.

    Second, not all taglibs are Gods gift to mankind. You don't *really* need a taglib that replaces if {} with <logic:if />, for {} with <logic:for ... /> and so on. It's more a matter of taste. And as you clearly prefer the PHP model where you embed program code inside the pages, I don't understand why you bother to use those taglibs in your JSP pages?

    Finally, if you don't like JSP, several different alternatives exist on the Java-platform. Actually I think that's both a big strength and weakness. You have a whole bunch of different ways to write webpages, cocoon, xmlc/barracuda, millstone, tapestry, just pick the one you like the most. As a result there is no typical way to write a Java web application, so when a comparison to another platform is to be made, one usually ends up comparing with JSP and perhaps Struts, being sort of least common denominator.
  175. Actually I can't believe this javashit vs othershit started all over again. It's at least childish how people would argue that <%= "Hello world" %> is more complex than the php equivalent and other would argue that TSS sucked until Howard got into the picture. Like Tapestry is build on anything but java. Don't you have something to do ?

    You can bet IBM is desilusioned with java since they are the authors of a java compiler that teaches you design patterns and best practices, and also of the J2EE app server nightmareSPHERE :-)) Huh !

    Actually let's not forget what IBM is doing. They make money. So if they can make money with java or with php they don't give a shit. They just make money.

    To Rolf: How about this: Develop a web application with PHP frontend with the business logic powered by Spring components on top of a Hibernate DAO layer, with Tangosol Coherence (so you make peace with Cameron) as 2nd level cache, the whole thing deployed on 200 linux boxes :-D.
  176. Java vs PHP??[ Go to top ]

    Yeah, it's like comparing apples to oranges. I've coded with several web scripting languages, including PHP, PERL, ASP, and seen them used for both small homesites and large enterprise-scale applications. PHP is nice and easy - but in what way would it be better than JSP pages used together with JSTL tag libraries (that are now standard part of J2EE platform)? If you absolutely want to have all your code on one layer, JSTL tags let you do database calls and xml handling, iteration, conditional statements - and there's no hint of Java scriptlets or try-catch exception handling anywhere.

    I think people get confused when they think Java is hard. I think things you can do with Java may be hard - but sometimes there simply aren't alternatives and often the alternatives would not be any easier.

    Want a simple web page? Start up tomcat, write some ugly scriptlet on HTML, differences between PHP or MODPERL or Java on JSP page are minimal.

    Want to do an enterprise system that has expected life-cycle of 20-80 years? Want to integrate with distributed transaction monitors, a few different databases, email services? Want to make it nice and portable, re-usable and maintainable? That's where the 'hard' design patterns and xml descriptors come to play.

    Guaranteed, full-blown J2EE environment is not easy for amateurs, not often easy for professionals either. But I don't think a scripting technique for web layer poses an alternative, save for a few special cases. J2EE could be easier - and trend seems to be to make it easier - but today for enterprise systems there's nothing I would rather use. (The question is more like what parts of it to use, and how to combine them - KISS is still a sound advice when you can use it and POJOs rule).
  177. I think the basis of whole discussion i wrong. The article quotes "IBM's been so fed up with Java that they've been looking for alternatives for years." So IBM want to throw out Java and rewrite Websphere in PHP? Do anybody find that even remote likely? That is simply ridiculous! :-()

    PHP might be the greatest web scripting language on the planet, but it is not a language designed to be used outside the web tier.

    You cannot compare Java (J2EE) and PHP, only in the context of web development.

    The whole idea of MVC is to separate the business logic from the data model and the view. With PHP you mix it together.
  178. I've seen PHP MVC.
  179. MVC[ Go to top ]

    I've seen PHP MVC.
    This stuff was available for CF and PHP before JSP 1.1 was released http://www.fusebox.org
  180. The controller part of MVC is of cause part of the web tier. What I mend was that the the controller should not contains business logic, only presentation logic, like fetching form data, call EJBs or what you use for your business logic, decide what view to return to the client. So my point was that business logic is outside of the web tier.
  181. I remember last year there was some discussion over the php integration in JSP. if I remember correctly, there was an alpha build. I'm just paraphrasing, but the report from those who tried it did experience bugs. I don't know what ever came of that, but maybe that's what IBM is planning. Just a thought.
  182. PHP documentation.[ Go to top ]

    Their docs site appears to be down
    http://za2.php.net/manual/en/print/index.php
  183. Good news[ Go to top ]

    I personaly think it is a good news. I wish IBM drop java completly and leave all JSRs.

    All IBM java tools and apps I have been forced to use so far were a pain.

    If java were free of IBM influence, then we could finally start to get real improvements from SUN and OS community.
  184. Good news -- Really???[ Go to top ]

    [...]
    If java were free of IBM influence, then we could finally start to get real improvements from SUN and OS community.

    Huhh??? You don't mean that, do you? In fact, IBM donated tons of code to OS communities (for example: Xerces, Log4J to the ASF, not to forget all the eclipse stuff).

    Regards,
        Dirk
  185. By now there are a lot of opinions in this thread. Generally I keep finding something that is in common:

    -PHP focuses on web development and has lots of supporting libraries.
    -PHP is easy to learn
    -PHP is good for quick scripting and it can be productive in small to midsize projects.
    -PHP can work well in the hosting environment, and definately in 'lamp' configurations.
    -PHP security is a bit of an issue.

    Is it better than Java then? Like people have pointed out, PHP only focuses on Web development, and in that role it is not directly comparable to Java.

    So we should be talking about Java web technologies vs. PHP. Now this is interesting because Java has many standard and non standard technologies to choose from. They all have their own niche, which can be combined, used separately, or what ever the need is. I'd say that Java aswers a far wider range of requirements than PHP in this sense.

    Web development is not just about rendering web pages though. This is where the comparison gets blurry. Take databases for example. Does PHP have a productivity edge here? I would definately say no. This is just one example but it shows that the overall judgement has to be made by application basis.

    The big picture is that Java can be the safer choice. While it might not match the productivity of PHP, efficiency in a hosting environment, on case by case basis it offers very competitive alternatives due to its large array of solutions to web development, and definately exceeds the range that PHP is capable of in many cases. Some of these Java technologies take longer to learn but there is no gain without pain.

    But then look at something like .NET which has a very nice Web development environment, but it can't bend over backwards like Java can to match many kinds of requirements, and it can't compete with PHP outside the Windows world. The anonymous analyst should take a jab at .NET.

    So take your pick!
  186. Don't take this news to hard. IBM is just trying to gain a bigger market share by endorsing PHP, just like Borland makes tools for every platform (Java, .NET) and even older languages such as Pascal.
    In December of 2004 NYSE (New York Stock Exchange) announced that it's rolling out a Java based trading system worth at between $100m and $200m and built by IBM to replace an older system written in C++.
    Please read Java for NYSE article. I quote here "Analyst Dave Cearley at Meta Group said the project marks a turning point for Java. The interesting part is not simply what this provides to the New York Stock Exchange, but what this means for the larger position in the market for Java. This is not simply a statement of the scalability of IBM's proprietary technologies. It is also about showing the scalability and reliability of Java environments."
    This proves to me without a doubt that IBM knows and believes that Java and J2EE are good enough to handle billions and billions of dollars; otherwise they would have used some other technology, right? :-)
  187. "IBM knows and believes that Java and J2EE are good enough to handle billions and billions of dollars"

    "NYSE has been under pressure from all-electronic exchanges such as the Nasdaq Stock Market, which can perform trades faster.

    Well I hope they will succeed! Nasdaq runs on .NET technology - I wonder how much they paid? 20% of NYSE perhaps?

    Oracle was dumped:
    Initially, this audit log was maintained simply through network I/O to parallel Oracle databases running on the Intel® Xeon™ platform. With a significant increase in the number of transactions occurring, NASDAQ realized that a more powerful system was required to ensure that less than 10ms latency was maintained. The resulting solution was made up of a mix of 4-way and 2-way Intel® Itanium® 2 boxes supporting a cluster of SQL databases.
    http://businessweek.itpapers.com/abstract.aspx?cid=71&dtid=3&docid=130178

    But they have to hurry up - already Nasdaq has new Microsoft tricks in the bag:

    http://accounting.smartpros.com/x34942.xml
    http://www.nasdaq.com/xbrl/

    I hope Peter Lin is reading this!

    Regards
    Rolf Tollerud
  188. .NET playing the catch up game[ Go to top ]

    Nasdaq runs on .NET technology - I wonder how much they paid?

    I see evidence that NASDAQ uses clustered Windows servers, and that NASDAQ provides some web services using .NET, such as automated SEC filings, but I can see nothing that implies that NASDAQ as a whole 'runs on .NET technology'. Do you have a URL?

    Their move to Itanium is interesting, as it seems to be a rather unsuccessful processor.

    Of course, J2EE is well established in providing very high-volume such as EBay, so it seems that it is .NET that is trying to catch up.
  189. finacial system on a budget[ Go to top ]

    Steve,

    "Do you have a URL?"

    Certainly,
    http://www.microsoft.com/resources/casestudies/CaseStudy.asp?CaseStudyID=14483
    Five months before the scheduled release of its mission-critical SuperMontage order display and execution system, NASDAQ concluded that the NASDAQ Prime data component it had built on a proprietary database would not meet the massive throughput and business-logic demands of the system

    They do not say what technology that failed! (you have to use your imagination)
    32,000 database calls per second in production and 60,000 database calls per second in testing build in 10 weeks.

    As I said - I wonder how much they paid?

    Regards
    Rolf Tollerud
  190. finacial system on a budget[ Go to top ]

    Steve,"Do you have a URL?"Certainly,http://www.microsoft.com/resources/casestudies/CaseStudy.asp?CaseStudyID=14483

    That is nothing to do with .NET. I asked for a URL to back your statement that 'NASDAQ runs on .NET technology'. There is not a single mention of .NET on that page.
    Five months before the scheduled release of its mission-critical SuperMontage order display and execution system, NASDAQ concluded that the NASDAQ Prime data component it had built on a proprietary database would not meet the massive throughput and business-logic demands of the system

    And the Microsoft solution they used was not a 'proprietary database'?
    They do not say what technology that failed! (you have to use your imagination) 32,000 database calls per second in production and 60,000 database calls per second in testing build in 10 weeks.
    As I said - I wonder how much they paid?RegardsRolf Tollerud

    Firstly, having taken a look at the details of the situation, no technology had 'failed': the system they were using had been working fine. They were simply looking at a migration to 64 bit, and needed to do it fast. The URL you gave points to a rather strangely written report obviously produced by consultants who were not IT literate - I quote: "a cluster of SQL databases rather than Oracle databases" (so Oracle is not an SQL database?). The migrated system was developed using stored procedures and specialised code, not .NET! The report said that speed of implementation was a key factor. I would agree that Microsoft systems can be set up quickly. Of course, an appropriately set up Oracle system would have worked as well: The Philadelphia Stock Exchange runs Oracle on Solaris, and handles 75,000 quotes per second (rather more than the 8,000 quotes per second of that NASDAQ system!). They are testing Solaris 10 and Linux + Oracle in order to cope with 120,000 quotes per second.

    So, no evidence that NASDAQ 'runs on .NET technology'.
  191. Steve,
    "SQL databases rather than Oracle databases"

    You mince words. They just lost the server word, it should be read, "SQL Server databases rather than Oracle databases".

    "The migrated system was developed using stored procedures and specialised code, not .NET!"

    If you browse the site you will find both .asp and .aspx extensions. That part the site was done in "Classic Asp" is even worse for the Sun/Java/Oracle camp.

    "The Philadelphia Stock Exchange runs Oracle on Solaris, and handles 75,000 quotes per second"

    It should be obvious by now that Microsoft technology can handle anything that Java can, and for a much lower price. Just let it take some time to sink in.

    No civilizations or technology lasts for ever. That is the lesson from history. During the transition phase from the old to the new, the old do everything to discredit and ridicule the new. Tell me something new, yawn..

    Look at the other thread with Ruby on Rails, listen to the outcries! A toy language! Yet when I look in the Crystal ball I see that the "next language" will be:

    1) Without strong typing
    2) Very sparse and economic
    3) Not compiled during the development, compiled in production

    If it's not Ruby it will be something else along these lines.

    Microsoft has infinitely more money and resources, as well as a better company culture than the opposition and is in an extremely well position to out compete anybody. The only thing that kept them back was that the "tiny PC processor" - that is not so tiny anymore. Consequently there are nothing that can protect IBM and the other "mainframe type" of companies from their outrageous overcharging.

    Don't be an army mule. If you know both Java and .NET you are in a very good position. Just look up IKVM/GNU Classpath and find out how to run classic Java on CLR - the best of both worlds.

    Regards
    Rolf Tollerud
  192. You mince words. They just lost the server word, it should be read, "SQL Server databases rather than Oracle databases"

    This was clearly illustrating the poor quality of the report. They were the ones literally mincing words! It is a silly as calling 'Windows Server' just 'Server'. If someone wrote in a public report 'In recent years people have migrated from Netware to Server' I hope you would gain a low impression of their technical competence.
    The migrated system was developed using stored procedures and specialised code, not .NET!"
    If you browse the site you will find both .asp and .aspx extensions. That part the site was done in "Classic Asp" is even worse for the Sun/Java/Oracle camp.

    This is known in debate terms as 'reaching'. You said 'NASDAQ runs on .NET technology'. You can provide no evidence other than some .aspx extensions on some pages. This shows that they use .NET for some web services. I already pointed that out. Like most large organisations, NASDAQ uses a range of technologies. They most certainly do not 'run on .NET technology'.
    "The Philadelphia Stock Exchange runs Oracle on Solaris, and handles 75,000 quotes per second"
    It should be obvious by now that Microsoft technology can handle anything that Java can, and for a much lower price.

    A classic sign of someone having lost an argument is when they finally resort to stating personal opinion as 'obvious' fact with no evidence.
    If you know both Java and .NET you are in a very good position. Just look up IKVM/GNU Classpath and find out how to run classic Java on CLR - the best of both worlds.RegardsRolf Tollerud

    Yes. If you look up 'Java' on www.sun.com you can find out how to run full modern Java on all platforms that CLR runs on and more. Even better, you can get it for free!
  193. here comes an other one[ Go to top ]

    More and more financial systems move to C#/.NET.

    Here is another link that will shatter Peter Lin.
    http://www.wstonline.com/story/WST20030620S0012
    With competition heating up in the direct-access space, SunGard Trading Systems/BRASS (is unveiling a new front-end interface called U2 from which broker/dealers and buy-side institutions can launch smart-order routing to multiple trading destinations.

    With a nod to the Irish rock band, U2 is being written in Microsoft C# and is based on the Microsoft .NET platform. U2 may eventually replace UMA - Universal Market Access - the trade-order-management vendor's existing market-access platform, which is developed in Unix and uses a Motif front-end.


    Observe how Tom King, president and Chief operating officer answer the question:
    "why didn't SunGard choose Java or XML for U2?"

    Regards
    Rolf Tollerud
  194. there went an other one[ Go to top ]

    More and more financial systems move to C#/.NET.Here is another link that will shatter Peter Lin.http://www.wstonline.com/story/WST20030620S0012
    With competition heating up in the direct-access space, SunGard Trading Systems/BRASS (is unveiling a new front-end interface called U2 from which broker/dealers and buy-side institutions can launch smart-order routing to multiple trading destinations. With a nod to the Irish rock band, U2 is being written in Microsoft C# and is based on the Microsoft .NET platform. U2 may eventually replace UMA - Universal Market Access - the trade-order-management vendor's existing market-access platform, which is developed in Unix and uses a Motif front-end.
    Observe how Tom King, president and Chief operating officer answer the question:"why didn't SunGard choose Java or XML for U2?"RegardsRolf Tollerud

    You seem not to understand statistics. One example from two years ago does not indicate a trend.

    Also, you picked a bad example to make your case. At SunGuard, the .NET platform has not taken over two years later. They are still using Unix front-ends, and have no timetable for any new Windows system taking over, and Tom King has left. Guess what server system they are using? Solaris!

    So, your best example of a 'financial system moving to .NET' is actually a financial system failing to move to .NET after many years even on the client side, and using Solaris as the server system?

    I can't really speak for anyone else, but I would not imagine that this will shatter anyone.
  195. Ralf: It should be obvious by now that Microsoft technology can handle anything that Java can, and for a much lower price. Just let it take some time to sink in.

    How much time does it need? Compared to JoNaS (jOnAs?) / JBoss / Geronimo with MySQL / PostGres running on Linux, the .NET platform with SQL Server looks pretty pricey. And for the low-end applications, how can you beat PHP / Python / Perl on Linux with Apache and MySQL / PostGres? And Microsoft isn't present on the high end, except for a few accounts that they paid to be references, such as NASDAQ. (Oh, was that supposed to be a secret? Maybe they could trot out Newport News again. Hmm .. I wonder if Gates is the largest shareholder or something?)

    Ralf: No civilizations or technology lasts for ever. That is the lesson from history.

    Exactly. That's what most of us saw starting to happen in the '95 / '96 timeframe, as Java caused Microsoft to lose its incomprehensible death grip on the industry. The result was an explosion in innovation.

    Now, 10 years later, Java is the old guard getting targeted by those new upstarts, the Ruby and PHP kiddies. I just can't believe that you are so far behind the curve.

    Ralf: Microsoft has infinitely more money and resources, as well as a better company culture than the opposition and is in an extremely well position to out compete anybody.

    It must really bother you to have to defend an historical anachronism such as Microsoft or IBM. You must have invested a lot of time in learning all of Microsoft's technologies; don't worry, it's not that hard to learn something new.

    Ralf: Don't be an army mule. If you know both Java and .NET you are in a very good position. Just look up IKVM/GNU Classpath ..

    Or, in the words of Gilbert (you should try to read more!):
    The idiot who praises, with enthusiastic tone,
    All centuries but this, and every country but his own.

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  196. Exactly. That's what most of us saw starting to happen in the '95 / '96 timeframe, as Java caused Microsoft to lose its incomprehensible death grip on the industry. The result was an explosion in innovation.

    Exactly. Those of us old enough to have followed Microsoft since they started realise that they have rarely if ever innovated - they have almost always been reactive and trend followers. Windows followed MacOS, Word and Excel followed Word Perfect and Lotus, and .NET was nothing more than a duplication of Java, but within Microsoft's licensing terms and OS limitations, after (thanks to Sun) Microsoft had failed to embrace, extend and limit Java.
    Now, 10 years later, Java is the old guard getting targeted by those new upstarts, the Ruby and PHP kiddies.

    I would actually disagree with you here! Scripting languages are old: PERL and TCL appeared a long time before Java. Also, I would argue that Java with useful speed and stability is relatively recent(I remember using Java when it first came out). I believe that Java, with its garbage collection, portability and highly tuned VM, is the innovative platform. The scripting language approach is old and has not progressed very far.
  197. the future belongs to Microsoft[ Go to top ]

    Links from 1998 from a company nobody has heard about or no links at all..

    I had 3 links about Nasdaq and still Steve question that Nasdaq runs on Microsoft technology. Cameron comes with links from 1998 from a company nobody has heard about..

    To begin with, please show some links about The Philadelphia Stock system that handles 75,000 quotes per second. The last time we had this discussion with Peter Lin, the system he referred to was run on mainframes.

    Let me also remind you that it never in any time in TSS history, (neither in anyone else history) has been shown that a Big EJB Java Application Elephant server can run a high volume mission critical site. Not even a single, 1 (one) time. Never!

    To say that Microsoft is only good for small system is so ridiculous that the argumentation falls back on itself. The fact was and still is: The Oracle was dumped because of low performance - the site now works with Microsoft technology. Also Oracle is routinely out competed at TPC (http://www.tpc.org/tpcc/results/tpcc_price_perf_results.asp).

    Now I sit back and await links from the Philadelphia Stock Exchange. Most likely there will not come any, they will claim "Non Disclosure Agreements" - they usually do. How come that this isn't a problem in any of the 600 case stories at http://www.microsoft.com/resources/casestudies/?

    Regards
    Rolf Tollerud
  198. P.S. don't bother[ Go to top ]

    Today PHLX (Philadelphia Stock. Exchange) not only has the IBM mainframe platform, but also servers from Stratus Technologies and SunMicrosystems, to interface between the end peripheral and the Stratus boxes.

    http://64.233.183.104/search?q=cache:XpC6xXQxlrsJ:www.consul.com/files/CaseStudies/US_Consul_PHLX.pdf+PHLX+ibm+mainframe&hl=sv
  199. the future belongs to Java[ Go to top ]

    Links from 1998 from a company nobody has heard about or no links at all..I had 3 links about Nasdaq and still Steve question that Nasdaq runs on Microsoft technology.

    I don't question that NASDAQ runs one system on Microsoft technology, using stored procedures on SQL Server (or 'SQL' as the report you quoted said). That is not what you said. You said that 'NASDAQ runs on . NET technology'. Trying to change what you said won't work.
    Cameron comes with links from 1998 from a company nobody has heard about..To begin with, please show some links about The Philadelphia Stock system that handles 75,000 quotes per second. The last time we had this discussion with Peter Lin, the system he referred to was run on mainframes.

    Here are a couple of sources:

    http://www.phlx.com/news/pr2005/05pr032105.htm
    http://www.wallstreetandtech.com/showArticle.jhtml?articleID=56900618
    Let me also remind you that it never in any time in TSS history, (neither in anyone else history) has been shown that a Big EJB Java Application Elephant server can run a high volume mission critical site. Not even a single, 1 (one) time. Never!

    Nonsense and you know it. EBay is a high volume mission critical site, and it runs on J2EE/EJB, on websphere - a 'Big EJB Java Application Elephant server'! So, you can't say this ever again.
    To say that Microsoft is only good for small system is so ridiculous that the argumentation falls back on itself.

    Not at all. In fact, NASDAQ was not confident that Microsoft systems could cope with their medium-scale system.
    The fact was and still is: The Oracle was dumped because of low performance

    No. It was not 'Oracle', it was the entire 32-bit system. Oracle was just one component of it. You could just as easily say that 'Intel 32-bit processors' were dumped! Obviously, Oracle can easily handle high volume systems. For example Oracle has been handling EA on-line gaming: For years, The Sims has used Oracle Real Application Clusters to support 35,000 simultaneous users and 30,000 SQL transactions per second.
    Oracle is routinely out competed at TPC (http://www.tpc.org/tpcc/results/tpcc_price_perf_results.asp).Now I sit back and await links from the Philadelphia Stock Exchange.

    No need to wait. Provided above.
  200. Why do you care?[ Go to top ]

    To say that Microsoft is only good for small system is so ridiculous that the argumentation falls back on itself.

    I have to finally ask - why do you care? Why the obsession with Microsoft everywhere? Surely a good IT manager/developer sticks to good general business practices and aims for technologies that allow competitive tendering for services. Any single company dominating any aspect of IT (even if they have done this through merit) is a bad thing for developers and users because it reduces competition. Isn't that plain and obvious? You don't need to stick to Microsoft. If you really can't stand Java, there is the cross-platform option of Mono (if you are prepared to take the chance that Microsoft wouldn't license them to death if they ever got too successful). Mono seems to be doing reasonably well, and should cope with most of your .NET work.

    Over decades I have dealt with very many computer companies. I would not trust any single company. My IT strategy is based on the ability to source everything I use from more than one vendor. I have a lot of trouble understanding why anyone would do anything else.

    I'm not after a long debate - I just would like to understand where you are coming from.
  201. sorry cant help myself[ Go to top ]

    Steve,

    "I have to finally ask - why do you care?"

    Good question - to tell the truth if it only was about standing up against a Big Company I wouldn't care.

    But the big mistake they (the J2EE crowd) did was to assume the self-imposed title of "Computer Scientists" that speak in big letters only, like no one ever done before in any computer language group I know of. Then I can not stop myself, after all I am audacious, rude and curious - not the kind of guy that take lightly to be spoken to in terms of "Organize inter-service transfers according to use cases from known domain objects into a coarse-grained Composite".

    BTW, "EBay is a high volume mission critical site, and it runs on J2EE/EJB"

    A couple of things are quite striking. The first is its complete lack of usage of Entity Beans

    But at least it runs on Java! ;)

    Regards
    Rolf Tollerud
  202. P.S. this is for Cameron[ Go to top ]

    The last is the stateless nature of the system and the conspicuous absence of clustering technologies.

    From the same link as above..
  203. Back to you Rolf[ Go to top ]

    The last is the stateless nature of the system and the conspicuous absence of clustering technologies.From the same link as above..

    More from that useful link:
    The impressive part is that eBay had 380M page views a day with a site availability of 99.92%. In addition to that, nearly 30K lines of code changes per week. Just plain and simply enviable, not only that, incontrovertible evidence of the scalability of Java.

    and
    use the O-R mapping to route to different data sources.

    Yes, one of the highest volume websites in the world uses ORM.
  204. P.S. this is for Cameron[ Go to top ]

    The last is the stateless nature of the system and the conspicuous absence of clustering technologies.

    From the same link as above..

    Just keep telling yourself that ;-)

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  205. The last is the stateless nature of the system and the conspicuous absence of clustering technologies.From the same link as above..
    Just keep telling yourself that ;-)Peace,Cameron PurdyTangosol, Inc.Coherence: Cluster your POJOs!

    Indeed. Note to self: always check Rolf's links.

    EBay does use clustering. After some serious system failures years ago they changed their policy and moved to Veritas clustering for their servers:
    http://www.snwonline.com/case_studies/ebay_03-28-05.asp?article_id=530

    The quotes in the article Rolf mentions are from JavaOne in 2000; nearly 5 years ago, when EBay was still using its old strategy.

    Rolf has obviously given up posting out-of-date links himself. He now provides links to others who do this for him.
  206. It is nothing wrong with clustering as long as you have what is described as basically a stateless system.

    It is when you combine stateful with clustering that you open the can of worms.

    Regards
    Rolf Tollerud
  207. How wrong can you be?[ Go to top ]

    It is nothing wrong with clustering as long as you have what is described as basically a stateless system. It is when you combine stateful with clustering that you open the can of worms.RegardsRolf Tollerud

    Ah, so now you agree that EBay uses clustering. They also use J2EE, they use Solaris servers and they use Oracle to support their high volumes of transactions.

    They also use EJBs - they use stateless Session Beans. An aspect of J2EE they don't use is Entity beans. Instead they use their own ORM.

    So, let's see. EBay uses Oracle, they use clustering, they use Solaris, and they use a 'Big Elephant J2EE server' (WebSphere) to run large parts of J2EE, including EJBs.

    So, you can never again quote nonsense about Oracle not being able to deal with high volumes, Java not being fast and scalable and J2EE/EJB not being used in high volume sites.
  208. let us sum it up[ Go to top ]

    1) Ebay is constructed very different from ordinary J2EE/EJB projects.
    2) They advocate statelessness as much as possible. (like yours truly :)
    3) They did not have any EJB when the system was new; nothing definite has been shown to prove that policy has changed.

    They choose Java precisely in the moment when .NET was new and had not yet proved itself. That was exact the reason "We are not betting on a new unproven technology". You can think what you want about that decision - but it was probably the biggest mistake in the history of computing.

    You can just imagine the costs to harbor the 100+ developers and having IBM in fulltime for all this years! You can not even count to so much..

    The amount of money that could have been saved with Microsoft technology boggles the mind.
     
    Regards
    Rolf Tollerud
  209. 1) Ebay is constructed very different from ordinary J2EE/EJB projects.

    Any project of that size will be constructed differently from a smaller project. Perhaps we could label this an 'extraordinary' J2EE/EJB project, to indicate how well it adapted to the demands.

    I note you now accept it is a J2EE/EJB project (proof below). I'm glad we are making progress.
    2) They advocate statelessness as much as possible. (like yours truly :)

    Irrelevant to the clustering/J2EE/EJB argument.
    3) They did not have any EJB when the system was new; nothing definite has been shown to prove that policy has changed.

    The original system was not clustered and was poorly designed. It failed. They re-implemented using a new strategy, including EJBs.

    As for 'nothing definite', you are wrong (yet) again: Here is a TSS statement from John Crupi, describing the implementation of EBay 3.0:
    "they're using stateless session beans".
    They choose Java precisely in the moment when .NET was new and had not yet proved itself. That was exact the reason "We are not betting on a new unproven technology".

    They still did not use .NET a couple of years later when they re-implemented the system. By that time .NET was as mature as J2EE was when they first set up. They were actually in negotiations with Steve Ballmer about using .NET for EBay. Microsoft had been after the EBay deal "for a long time". EBay They still chose J2EE over .NET. They were willing to use .NET, but it lost out on EBay's internal tests.
    You can think what you want about that decision - but it was probably the biggest mistake in the history of computing.

    Woo! You are saying that EBay's use of Java is "probably the biggest mistake in the history of computing"?

    That has to be one of the nuttiest things I have ever seen posted on a supposedly sensible site. It is obviously wrong. Microsoft's lack of security is a bigger mistake - Windows viruses it has resulted in proven costs to the industry of billions (e.g. http://www.infoworld.com/articles/hn/xml/00/05/08/000508hnlovehurts.html).
    Whereas EBay is in profit, and that profit is increasing.

    So, this is obviously not 'the biggest mistake in the history of computing'. You are provably wrong again. This is getting boring.
    You can just imagine the costs to harbor the 100+ developers and having IBM in fulltime for all this years! You can not even count to so much..The amount of money that could have been saved with Microsoft technology boggles the mind.&nbsp;RegardsRolf Tollerud

    Yes, you can always save money initially by using cheap and innappropriate technology.

    Tell you what - why don't you contact EBay and tell them all this? You obviously seem to know far more about IT than they do.

    Mind you, even Ballmer can make mistakes, so don't feel bad. He said that EBay had made the wrong decision and would change their minds. That was 4 years ago, and EBay is one of the most successful sites on the Web.
  210. Cameron is a real tough hardliner[ Go to top ]

    We are not getting anywhere. This is exactly the same bickering I had with Cameron three years ago. Since that the following has happened:

    1) EJB has lost out big
    2) Spring and the other IOC systems have appeared and takes marketshare.
    3) IBM has started to support PHP.
    4) Scripting languages as Python and Ruby are on the move, Java's Tiobe index has fallen from 24.24 to 16.98, an incredible 30% loss in one year.
    5) AJAX is contemplated for serious projects.
    6) Rich Clients are used more than for 3 years ago.
    7) SOA is big; the Sun/Java camp has nothing like Indigo.
    8) Declarative programming is big; the Sun/Java camp has nothing like XAML.
    9) .NET sites are more common that Java sites, according to Netcraft.
    10) Mono has been released for Linux - they are talking about building the next "Java desktop" in Mono! (what can I say ;)

    All this are taking marketshare from J2EE/EJB and even from Java itself. Still Cameron admits nothing.

    Regards
    Rolf Tollerud
  211. Here we go again[ Go to top ]

    We are not getting anywhere. This is exactly the same bickering I had with Cameron three years ago. Since that the following has happened:1) EJB has lost out big2) Spring and the other IOC systems have appeared and takes marketshare.3) IBM has started to support PHP.4) Scripting languages as Python and Ruby are on the move, Java's Tiobe index has fallen from 24.24 to 16.98, an incredible 30% loss in one year.5) AJAX is contemplated for serious projects.6) Rich Clients are used more than for 3 years ago.7) SOA is big; the Sun/Java camp has nothing like Indigo.8) Declarative programming is big; the Sun/Java camp has nothing like XAML.9) .NET sites are more common that Java sites, according to Netcraft.10) Mono has been released for Linux - they are talking about building the next "Java desktop" in Mono! (what can I say ;)All this are taking marketshare from J2EE/EJB and even from Java itself. Still Cameron admits nothing.RegardsRolf Tollerud

    Well, I can see why. The reason is that you provide no proof. You simply state things as fact. I understand why you provide no proof. When you do, it is so easy to reveal how wrong you are when you try to. As in the .NET site that wasn't, and the supposed absence of 'Big J2EE Websites', the 'non-clustered' website that was clustered. Every time you try and prove things you help provide great evidence to the contrary!

    You can't even find a single site that makes a consistent case: The same TIOBE that 'shows' Java 'falling' also shows C# hardly used by anyone!

    You are like a stuck record, repeating the same falsehoods, and presenting no evidence.
  212. more wrongess[ Go to top ]

    1) EJB has lost out big

    Exactly the reverse if you check the job market. EJB, whatever you think of it, is hugely in demand.
    2) Spring and the other IOC systems have appeared and takes marketshare.

    Yes. This shows that Java is active and exciting nd evolving and by no standards a 'legacy' language.
    3) IBM has started to support PHP.

    IBM supports everything. It still supports Smalltalk. This has no relevance to its strong and growing Java support.
    4) Scripting languages as Python and Ruby are on the move,

    Hardly at all, according to job searches. Ruby is small and static. I would agree that Python is getting interesting. Of course, both integrate cleanly with Java - JRuby and Jython. They are not a threat, they extend what the Java platform can do.
    Java's Tiobe index has fallen from 24.24 to 16.98, an incredible 30% loss in one year.

    Assuming that the Tiobe index has any meaning in terms of language popularity or use. If you do, then .NET is hardly used at all.
    5) AJAX is contemplated for serious projects.

    But not for real external websites, where there is a wide range of browsers and levels of JavaScript support.
    6) Rich Clients are used more than for 3 years ago.

    But still hardly at all.
    7) SOA is big

    And mostly implemented in Java. Java is the dominant service provider language by far.
     
    the Sun/Java camp has nothing like Indigo.

    Yes it does. There are plenty of transparent transactional messaging and remoting services. You can combine Spring with RMI or SOAP and transactions to publish and use distributed services very easily.
    8) Declarative programming is big

    Ah, like the use of annotations in Java 1.5, or the Jakarta commons annotations before that.
    the Sun/Java camp has nothing like XAML.

    I guess you have not heard of JSF, which is also an object/component based page rendering system. Multi-vendor, cross-platform, open source and, works with websites, WML, SVG etc. There is also Tapestry etc.
    9).NET sites are more common that Java sites, according to Netcraft.

    Er no. Sites that have pages with .NET type extensions are more common than Sites which have pages and URLs that can easily be detected as .JSPs or servlets. This is not the same thing at all.
    10) Mono has been released for Linux - they are talking about building the next "Java desktop" in Mono! (what can I say ;)

    I assume you mean Novell's Linux Desktop, which has precisely two (2!) significant Mono applications.
    All this are taking marketshare from J2EE/EJB and even from Java itself.

    Not according to the job markets. Java is finally overtaking C++ to become the most in-demand language.
  213. the java job market?[ Go to top ]

    Steve,
    "Not according to the job markets. Java is finally overtaking C++ to become the most in-demand language."

    Figures from UK's JobServe (it.jobserve.com/)
    For the past 5 days according to Yann Caroff
    http://www.theserverside.com/news/thread.tss?thread_id=17627&article_count=44

    In January 29, 2003:
    ,NET was 32.6 % of Java

    In March 9 2005:
    ,NET was 61.8 % of Java

    And when I check right now (one month later)
    ,NET was 62.7 % of Java (up 0.9 % in one month)

    So according to this .NET will overtake Java in about three year from now.

    Of course that means that Java will have lost its significance long time before that as the Java jobs are mostly old maintenance jobs while the .NET jobs are new projects.

    Regards
    Rolf Tollerud
  214. the java job market?[ Go to top ]

    Steve,"Not according to the job markets. Java is finally overtaking C++ to become the most in-demand language."Figures from UK's JobServe (it.jobserve.com/) For the past 5 days according to Yann Caroff http://www.theserverside.com/news/thread.tss?thread_id=17627&article_count=44In January 29, 2003:,NET was 32.6 % of JavaIn March 9 2005:,NET was 61.8 % of JavaAnd when I check right now (one month later),NET was 62.7 % of Java (up 0.9 % in one month)So according to this .NET will overtake Java in about three year from now.Of course that means that Java will have lost its significance long time before that as the Java jobs are mostly old maintenance jobs while the .NET jobs are new projects.RegardsRolf Tollerud
    Rolf,
      You've got to be pulling our legs. There is no way any semi-delusional person would come up with that conclusion. Good statistics are not to be trusted let alone "statistics" with 3 data points (and one of them we would have to trust you about - riiiight).

    Even if you did trend it out
    - there are more variables than what you've stated
    - it would only prove advertised Jobs on that job board were changing percentages.

    3 years from now? Hmm if I trend out Microsoft dumping platforms - that is about the right time for them to orphan some more developers.
  215. this is fun![ Go to top ]

    Steve,
    "You've got to be pulling our legs. There is no way any semi-delusional person would come up with that conclusion."

    Yes?

    "Good statistics are not to be trusted let alone "statistics" with 3 data points (and one of them we would have to trust you about - riiiight)."

    1) For the first result I gave the link: http://www.theserverside.com/news/thread.tss?thread_id=17627&article_count=44

    2) For the second result check the thread ".NET expert: Microsoft is losing confidence in .NET"
    , search on Yann.

    3) For the third result just go to it.jobserve.com/, set the search to the last 5 days.

    I save the results every month and have done that since the beginning when Caroff first published the search. Why don't you do the same?

    Contrary to popular opinion my links and information are accurate. The trick Cameron and others do is to focus on some irrelevant part such as "which one of MS’s particular technology was it:" The database SQL Server, or classic ASP, or .NET?" when the real issue was: MS is winning. Or "you are against clustering" (what a silly thing to say) when it is the combination of cluster and Statefullness that was my intention to bash.

    It is a English word for "deliberate misunderstanding" that just now eludes my mind, Cameron’s favorite tactic: Deliberate misunderstanding and focus on irrelevant detail.

    Nevertheless, there is one more reason to that I do it which I forgot to mention:

    It's so damn entertaining!

    Regards
    Rolf Tollerud
  216. this is fun![ Go to top ]

    Steve,"You've got to be pulling our legs. There is no way any semi-delusional person would come up with that conclusion." Yes?"Good statistics are not to be trusted let alone "statistics" with 3 data points (and one of them we would have to trust you about - riiiight)."

    I think you were replying to Mike there..

    However, you rarely come up with even one statistic.
    1) For the first result I gave the link: http://www.theserverside.com/news/thread.tss?thread_id=17627&article_count=44

    You are doing your usual nonsense of taking one story from two years ago and extrapolating.
    2) For the second result check the thread ".NET expert: Microsoft is losing confidence in .NET" search on Yann.3)

    Those figures show nothing to back you up, as I clearly pointed out in that thread. Those figures show Java jobs growing healthily, and .NET Jobs growing too, but there is no indication that either is taking market share from the other. A lot of .NET is taking market share from Visual Basic!
     
    For the third result just go to it.jobserve.com/, set the search to the last 5 days.I save the results every month and have done that since the beginning when Caroff first published the search. Why don't you do the same?

    Because the figures don't mean what you are trying to make them mean.

    I have a metaphor that might help you understand. I call it 'the truck and the bicycle':

    Imagine there is a truck-maker who is very successful, and you see his trucks everywhere. Then along comes Mr Bill the Bicycle Maker with his LittleSoft bikes. Bicycles are popular and thousands get made. Mr Bill gets very excited and starts shouting about how successful he is in the TI 'Transport Infrastructure' market - 'I am outselling the truck makers'! Bicycle fanboys start to proclaim bicycles as the new transport revolution. They tie lots of bikes together with rope and show how the resulting tangle of wheels can carry pretty heavy things. Some people start to use bikes for carrying such heavy things. It can work sometimes, but pedalling uphill is hard.

    Meanwhile, the lorries roll serenely past.

    I find this a useful metaphor. For bikes read Ruby, Lorries Java, or for bikes read Windows, for Lorries UNIX.

    .NET is a rather big bicycle, with lots of wheels, but has the problem of only being able to go on a few roads.
    Contrary to popular opinion my links and information are accurate.

    They are often accurate but irrelevant. Like the time you showed that Windows servers were popular a couple of years ago in the Pacific region. Of course that has no bearing on today's market. You perform meaningless extrapolations.
    The trick Cameron and others do is to focus on some irrelevant part such as "which one of MS’s particular technology was it:" The database SQL Server, or classic ASP, or .NET?" when the real issue was: MS is winning.

    No - YOU focus on a technology. For example YOU said 'NASDAQ runs on .NET technology'. When you are proved wrong you try and wriggle out of it, by saying 'well I really meant Microsoft technology'. MS is not winning. That was (*sigh* yet again) a single sample. Heard of the terms 'statistical outliers'?
    Or "you are against clustering" (what a silly thing to say) when it is the combination of cluster and Statefullness that was my intention to bash.It is a English word for "deliberate misunderstanding" that just now eludes my mind, Cameron’s favorite tactic: Deliberate misunderstanding and focus on irrelevant detail.

    We can only debate what you actually write. We don't have mind-reading abilities. If you can't clearly write what you are thinking, that is not our fault.
  217. this is fun![ Go to top ]

    I think you were replying to Mike there..
    :) You are old. My wife's grandfather use to call me Mike. I never bothered to correct him. No point in it.
  218. this is fun![ Go to top ]

    I think you were replying to Mike there..
    :) You are old. My wife's grandfather use to call me Mike. I never bothered to correct him. No point in it.

    Oh dear. Yes, I am old! Well, you kids would probably think I was.
  219. this is fun![ Go to top ]

    I'm no spring chicken either.
  220. this is fun![ Go to top ]

    Steve,

    First, it wasn't Steve.
    The trick Cameron and others do is to focus on some irrelevant part

    Second, that is what you do. I'm sure your stats on this are correct. But they don't prove what you are saying. It is non sequitur ("An inference or conclusion that does not follow from the premises or evidence. A statement that does not follow logically from what preceded it." )

    Your conclusion could come true. It could be true now(but it isn't). But your conclusion is not proven by your evidence.

    My pancakes are round and flat. The world seems to be round so it must be flat too. (Or will be in 3 years. :) )
    Nevertheless, there is one more reason to that I do it which I forgot to mention:

    It's so damn entertaining!
    Yes, you definitely are a comedy of errors. I hope you don't take yourself and/or your conclusions seriously.
  221. this is fun![ Go to top ]

    But your conclusion is not proven by your evidence. My pancakes are round and flat. The world seems to be round so it must be flat too. (Or will be in 3 years. :) )

    I think it's more like this:

    "Let me quote from a blog that has a link to a site by the Flat Earth society that quotes from a 100-year-old book that says that someone found a perfectly flat area in their garden. That proves that all the Earth is flat. Everyone knows that flatness is winning."

    Then, when proved wrong:

    "All I was arguing was that the world was a certain shape. You are deliberately misunderstanding me!"

    (Sorry, but I enjoy this too :)
  222. that is no toy. Sorry![ Go to top ]

    "A mix of dual or quad processor servers, with 4 Gigs of RAM and 64bit CPU supporting a cluster of SQL databases, running Windows Server 2003 Advanced Edition."

    Fun to see a bunch of grown men presumable knowledgeable and reasonable intelligent trying to convince each other otherwise..

    That makes you better understand the curious fenomen that everytime you see a thoroughly lousy film, real shit like for instance "Catwoman" with Halle Berry, you always hear the same story, "During the time and place the film was shot "the mood and atmosphere was so fantastic - everybody was sure that the film was going to be a bigsuccess".

    Apparently the human being's capacity to lure himself is infinite.

    Regards
    Rolf Tollerud
  223. that is no toy. Sorry![ Go to top ]

    "A mix of dual or quad processor servers, with 4 Gigs of RAM and 64bit CPU supporting a cluster of SQL databases, running Windows Server 2003 Advanced Edition."
    Fun to see a bunch of grown men presumable knowledgeable and reasonable intelligent trying to convince each other otherwise..That makes you better understand the curious fenomen that everytime you see a thoroughly lousy film, real shit like for instance "Catwoman" with Halle Berry, you always hear the same story, "During the time and place the film was shot "the mood and atmosphere was so fantastic - everybody was sure that the film was going to be a bigsuccess".Apparently the human being's capacity to lure himself is infinite.RegardsRolf Tollerud
    Remember that when you point at others you are actually pointing 3 fingers at yourself.

    "Either how canst thou say to thy brother, Brother, let me pull out the mote that is in thine eye, when thou thyself beholdest not the beam that is in thine own eye? Thou hypocrite, cast out first the beam out of thine own eye, and then shalt thou see clearly to pull out the mote that is in thy brother's eye."
    Luke 6:42
  224. The lip of truth shall be established for ever: but a lying
    tongue is but for a moment.
    The Proverbs 12:19

    Truth shall spring out of the earth; and righteousness shall look down from heaven. The Book of Psalms 85:11

    Regards
    Rolf Tollerud
  225. Also,[ Go to top ]

    You shall not follow a crowd to do evil; neither shall you testify in court to side with a multitude to pervert justice.
    Exodus 23:2
  226. Also,[ Go to top ]

    You shall not follow a crowd to do evil; neither shall you testify in court to side with a multitude to pervert justice.Exodus 23:2
    Refer to my privious post. Need tweezers or pliers?
  227. You are no Ali.[ Go to top ]

    The lip of truth shall be established for ever: but a lyingtongue is but for a moment.The Proverbs 12:19 Truth shall spring out of the earth; and righteousness shall look down from heaven. The Book of Psalms 85:11 RegardsRolf Tollerud

    You can't even find applicable Bible quotes. Again, they are true, but don't apply ... sort of like this -

    "And he [Judas] cast down the pieces of silver in the temple, and departed, and went and hanged himself" Matt 27:5

    and

    "... Go, and do thou likewise." Luke 10:37
  228. ?[ Go to top ]

    "You are no Ali"

    Ali? What has he to do with a citation from Oscar Wilde???
  229. ?[ Go to top ]

    "You are no Ali"Ali? What has he to do with a citation from Oscar Wilde???
    Silly bookworm. Do a google on "Float like a butterfly, sting like a bee". See what comes up first.
  230. Muhammad Ali = who is he?[ Go to top ]

    No need to call names! (hurt)

    I will be right in the end anyway.
  231. Muhammad Ali = who is he?[ Go to top ]

    No need to call names! (hurt)

    I will be right in the end anyway.

    .. takes marbles, sulks away ..
  232. Muhammad Ali = who is he?[ Go to top ]

    Muhammad Ali = who is he?

    Me hopeth thou joketh.
    No need to call names! (hurt)
    Not meaning to call you names (out loud at least). It was more like "Silly rabbit, Trix are for kids".
    I will be right in the end anyway.
    Ahhh. "Pride goeth before the fall" - You can search book/chapter/verse yourself.
  233. Toy Story[ Go to top ]

    "A mix of dual or quad processor servers, with 4 Gigs of RAM and 64bit CPU supporting a cluster of SQL databases, running Windows Server 2003 Advanced edition."

    Ah, so the appropriate reaction is "Wow! Someone has used Windows for something. It must be taking over the world!".

    Let me remind you about the principle of the Gaussian distribution in statistics. This means that one or two samples of Windows being used for relatively high performance applications does not indicate where Windows is mostly used. It does not indicate the modal use. Just because Windows can be stretched to its limit (after all, NASDAQ themselves doubted it was capabile) in a few cases does not mean that this is a situation to which it is suited. You need to present a large number of verifiable samples, not one or two stories.

    But, of course, you know this, and you are just trying to convince yourself, for unknown reasons, that Windows is competing in the enterprise server space.


    Also, I'm afraid that no matter how hard I look I STILL can't see a .NET label on any of that. Am I missing something, or were you wrong?
  234. the elusive J2EE case[ Go to top ]

    I have ceased to respond, usually the best thing to do after personal attacks but you still go on. Ok.

    "Let me remind you about the principle of the Gaussian distribution in statistics. This means that one or two samples of Windows being used for relatively high performance applications does not indicate where Windows is mostly used."

    And the 600 case stories at:
    http://www.microsoft.com/resources/casestudies/ ?

    And about this fictive high performance, mission critical J2EE site that is so common why do we not have a look. I am not after the stellar 75000 trans per second in the mainframe Philadelphia system neither do I think EBay (the most expensive project in IT history) has anything to contribute to the debate.

    All I want is i single example on a modest 6000-8000 tps project done for ca 5-6 mill dollar, a run-of-the-mill project according to the J2EE crowd, such as Peter Lin compare to a tree in Amazon: "if a tree falls in the Amazon, but no one hears it".
    If you can show me only 1 (one) such case, thoroughly documented with emphasis on EJB use, I promise I never mention this again.

    Regards
    Rolf Tollerud
    (the little child in "The Emperor new Clothes")
    P.S.
    Do you know that this year is H.C Andersen 200 year anniversary?
  235. P.S.[ Go to top ]

    Just to refresh your memory.

    http://www.andersen.sdu.dk/vaerk/hersholt/TheEmperorsNewClothes.html
  236. the elusive J2EE case[ Go to top ]

    And the 600 case stories at:
    http://www.microsoft.com/resources/casestudies/ ?

    Independent stories please. I want links from the users, not from the vendors.
    neither do I think EBay (the most expensive project in IT history) has anything to contribute to the debate.

    The most expensive project in IT history? What ARE you ranting about? Please supply proof of this statement.

    EBay contributes to the debate because it proved you wrong.
    If you can show me only 1 (one) such case, thoroughly documented with emphasis on EJB use, I promise I never mention this again.

    And it proves you wrong again.

    EBay's use of J2EE is thoroughly documented (on TSS, and elsewhere), the transaction rate meets (and far exceeds) your requirements, and the J2EE installation cost was (as usual) not precisely given, but "several million dollars", and I have already provided documented evidence directly from an engineer involved that it involved EJBs (Session Beans).

    So it is now up to you to keep your promise.
  237. the word is out..[ Go to top ]

    It is so pathetique. For the first you had no showed me any proof or link that they are using EJB, for the second I wonder what a project of more than 2 billion dollar have to do with the discussion.

    You know something else? As I said before I have been socializing with some BSD people lately. Here the other day a guy I never met before on his own behalf (I had not said anything I promise) brought up the subject. "Do you know that there is not a single example" he said, referring to high load Big J2EE server and EJB.

    hi hi

    Regards
    Rolf Tollerud
  238. Keep your promise, Rolf[ Go to top ]

    It is so pathetique. For the first you had no showed me any proof or link that they are using EJB,

    I gave you the link. EBay's use of EJB Session beans is documented on this website, in the Tech Talk by John Crupi. Here is the quote:

    "Q. For those J2EE use cases how did they [EBay] distribute the topology of their application?

    A. So, part of the mandate of EJB is to be stateless. So basically, they're using stateless session beans,..."
    for the second I wonder what a project of more than 2 billion dollar have to do with the discussion.You know something else?

    Can't you read? I said the cost of the J2EE was several MILLION dollars.
    As I said before I have been socializing with some BSD people lately. Here the other day a guy I never met before on his own behalf (I had not said anything I promise) brought up the subject. "Do you know that there is not a single example" he said, referring to high load Big J2EE server and EJB.hi hiRegardsRolf Tollerud

    You have already agreed that EBay is a high load Big J2EE/EJB server:
    1) Ebay is constructed very different from ordinary J2EE/EJB projects.

    You did not say it was NOT an J2EE/EJB project, you said it was not an ORDINARY J2EE/EJB project.

    I disagree with your analysis of how ordinary, or otherwise the project is, but your statement clearly agrees that it is a J2EE/EJB project.

    So, it is J2EE, meets your transaction requirements, contains EJBs (the link above proves it) and was set up at a cost of no more than several million dollars.

    Keep your promise.
  239. end of discussion[ Go to top ]

    You did not give the link but I take it that it was this one,

    Interview with John Crupi - Chief Java architect, Sun Services Group
    http://www.theserverside.com/talks/videos/JohnCrupi/interview.tss?bandwidth=dsl

    This basically is a piece of PR from Sun. How come that the 600 case studies from Microsoft were not accepted from you because "I want links from the users, not from the vendors" but then you put forward this Sun crap as bonafida information?

    And in the paragraph precisely before the one with your quote he says:

    "So, two or three of the use cases I believe were implemented in J2EE, and they could not afford to have any performance hits with J2EE. And actually those use cases are working perfectly and now they're looking at implementing more use cases in J2EE."

    2 or 3 use cases! "I believe"! ;)

    I am sick and tired of discussing Java, the biggest hoax/con job in the history of IT business. It is no need. People everywhere are on to it, different solutions is coming up and is tested every day. For my part I rather play with Rails which I downloaded some minutes ago.

    It is not difficult at all today to see Java for what it is, there are so much to compare with, (.NET among else). But it was much more difficult for 3 years ago and I am a little proud of that I so early saw it.

    How did I do it? It was not difficult. It was their quasi scientific talk. Early in life I learnt that people that talks like this always are fakes, and that was right of course.

    Regards
    Rolf Tollerud
  240. You did not give the link but I take it that it was this one, Interview with John Crupi - Chief Java architect, Sun Services Grouphttp://www.theserverside.com/talks/videos/JohnCrupi/interview.tss?bandwidth=dslThis basically is a piece of PR from Sun.

    This is information published by The Server Side, not Sun.
    How come that the 600 case studies from Microsoft were not accepted from you because "I want links from the users, not from the vendors" but then you put forward this Sun crap as bonafida information?

    Because the information was published by The Server Side, not Sun. Are you saying that The Server Side published incorrect information in this case?

    Oh, you may be interested that EBay confirm this use of EJB. In 2003, at JavaOne, there was a discussion of their use if J2EE. one of the presenters was a system developer/architect from EBay who worked on the J2EE project.

      
    "So, two or three of the use cases I believe were implemented in J2EE, and they could not afford to have any performance hits with J2EE. And actually those use cases are working perfectly and now they're looking at implementing more use cases in J2EE."2 or 3 use cases!

    Yes 2 or 3 use cases at EBay. Those 2-3 use cases cover most of EBay's transactions.
    I am sick and tired of discussing Java

    Well stop it then. If you aren't going to stop, either

    (1) provide evidence that The Server Side has published incorrect information in this case, and also that EBay (the customer, not a J2EE vendor) is also publishing incorrect information.

    or

    (2) Keep your promise.

    Your choice: which is it to be?
  241. end of discussion[ Go to top ]

    How did I do it? It was not difficult. It was their quasi scientific talk. Early in life I learnt that people that talks [sic] like this always are fakes, and that was right of course.

    So how did we know he was just a windbag? It was not difficult. It was his quasi literary talk. Earl in life we learnt that people who talk like that are always fakes, and we were proven correct, of course.

    Bis interimitur qui suis armis perit ..

    (Oooh, I knew that phrase would come in handy some day!)

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  242. end of discussion[ Go to top ]

    How did I do it? It was not difficult. It was their quasi scientific talk. Early in life I learnt that people that talks [sic] like this always are fakes, and that was right of course.
    So how did we know he was just a windbag? It was not difficult. It was his quasi literary talk. Earl in life we learnt that people who talk like that are always fakes, and we were proven correct, of course.Bis interimitur qui suis armis perit ..(Oooh, I knew that phrase would come in handy some day!)Peace,Cameron PurdyTangosol, Inc.Coherence: Cluster your POJOs!
    Darn. You beat me to it. But definitely better than I would have done.

    I can't understand how Rolf thinks Java is so expensive, etc. I am saving my current company tons of money by using Java and Open Source. And I am doing OLAP and soon to be a Portal and ... . And for >90% I can go with another vendor if the need arises.

    I think he needs to visit a psychiatrist and get some issues worked out. Please, for the good of the TSS community, seek help.
  243. end of discussion[ Go to top ]

    Darn. You beat me to it. But definitely better than I would have done.

    I assumed it was good, but not having the benefits of a classical education, and not finding Latin on Babelfish, I could only wonder..
    Please, for the good of the TSS community, seek help.

    Well, I think he is useful. I find that battling against 'Rolf the Uber-Troll' is a good education and entertainment.
  244. end of discussion[ Go to top ]

    .. not having the benefits of a classical education, and not finding Latin on Babelfish, I could only wonder..

    Basically it means that it is doubly insulting to be stabbed with one's own sword. (When in doubt, Google: It is cheaper and easier than a classical edumacation, and has a better memory to boot.)

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  245. end of discussion[ Go to top ]

    Please, for the good of the TSS community, seek help.
    Well, I think he is useful. I find that battling against 'Rolf the Uber-Troll' is a good education and entertainment.
    True. Then at least for his own well being.

    Now to finish out this week.
  246. Waiting for proof[ Go to top ]

    Incidentally, I am still waiting for evidence that
    Nasdaq runs on .NET technology
    and that EBay is
    the most expensive project in IT history
    .
  247. It is nothing wrong with clustering as long as you have what is described as basically a stateless system. It is when you combine stateful with clustering that you open the can of worms.

    80 million page views a day can't be wrong ;-)

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  248. But the big mistake they (the J2EE crowd) did was to assume the self-imposed title of "Computer Scientists" that speak in big letters only, like no one ever done before in any computer language group I know of.

    You keep confusing J2EE with EJB. There is no 'J2EE crowd'. There is a huge and varied community, which use different aspects of J2EE.

    Guess what? I am, part time, a scientist. I would far rather use something that was designed by someone who knew what they were talking about and could prove it rather than some hacked cool solution. If you use a plane, or a large building, would you be happy if it was not designed by people who understood science? Why should IT projects be any different? "Scientist" is a positive term.
     
    Then I can not stop myself, after all I am audacious, rude and curious - not the kind of guy that take lightly to be spoken to in terms of "Organize inter-service transfers according to use cases from known domain objects into a coarse-grained Composite".

    You know what? I partly agree with you. I think that far too much obscure terminology has crept into IT. However, when you are developing large systems, it is sometimes necessary.
    BTW, "EBay is a high volume mission critical site, and it runs on J2EE/EJB" A couple of things are quite striking. The first is its complete lack of usage of Entity BeansBut at least it runs on Java! ;)RegardsRolf Tollerud

    No, it runs on... Java/J2EE. You forgot the J2EE kit. They also supply lots of toolkits for users to develop applications to use EBay functionality. Guess what is in some of those toolkits? EJBs.

    Also, check for job ads from EBay. There is one posted today on yahoo, which includes requirements for J2EE and EJB. I guess EBay does use EJBs in some of its software then, or is planning to. They are obviously not afraid of 'Science'.