Discussions

News: Ask TheServerSide: What is the state of Java?

  1. Ask TheServerSide: What is the state of Java? (159 messages)

    The O'Reilly editors have been discussing the state of Java.

    They are discussing: Is the Java revolution over? Has the big business-friendly approach turned Java into the new COBOL? Is the momentum behind .NET and Mono stealing away the cool factor that Java once enjoyed? Are those even bad things? What do you think?

    Read: The State of Java

    Threaded Messages (159)

  2. Which Java?[ Go to top ]

    Sun's "Java" the brand?
    Read this: http://www.onjava.com/pub/a/onjava/2004/10/06/jds.html
    They can't help themselfs(look at their Stock price), so they can't help others. Sun's Java in this case is a C desktop, and that is all they think of it.

    IBM's Eclipse JVM or BEA's JROckit VM are just fine, and GNU is cathing up. (They can't call it Java I think)
    The "Langage" formerly known as Java, is not keeping up on the server side, loosing to PHP. There are several companies that dumed Java (ActionWatch, etc.) to go to PHP. Look at the number of applications that are easy to install in PHP (Bugs, Content, Forums, etc.) Plus PHP version 5 has SOAP built in.
    So server side main competion is PHP and boy they are good.

    Java's main advanage is JNLP, the way to web deploy applications that bypass the browser (ex JDNC), when used with SoA such as Hessian.
    So Java's future growth is back on the Client side, and it will continue to shrink on the server (Unless Fujitsu KILLS Sun)

    Mono looks STRONG!!!!!! (JUnit, Ant, Ibatis, Spring all ported). And Mono is cross platform, Mac, Windows, Linux.

    Java "the brand" was killed dead by Sun's EJB marketing. KISS allways wins out. In J2SE 5.0, Java jumped the shark by not offering enough Client side, where it has an advantage.

    (Still time for IBM to expand things like NetX or JNLP, but.. they are stuck on their RCP, not on operations and deployment, where NetX is and the need is)

    How do I web deploy Mono anyone?
    .V
  3. They're leaving a lot out[ Go to top ]

    This critique feels like it's only talking about official java apis approved by JSRs and Sun. There is a lot of really interesting cutting edge stuff happening in Java in the open source community. Take a look at all the ground breaking O/R mapping packages, all the various web frameworks, lightweight containers, aspect oriented programming projects, etc.

    If you focus on EJB and commercial app servers, Java looks a lot like COBOL, but that isn't the whole story.


    Having programmed with a variety of scripting languages (PHP, Perl, pre-java Cold Fusion), I don't agree that the scripting languages are encroaching on Java. There is a different audience using these languages and while you can make a good website in PHP (and very quickly), you're not going to create a maintainable, portlable, extensible application. It's just not the same thing. Because of that, once you approach a certain scope, Java is a much better solution. PHP 5 is a step in the right direction, but Java has an incredible edge.

    On the topic of .Net and Mono I dunno. .Net, so long as it is confined to windows only is a fairly uninteresting. Mono maybe be good stuff, but at this point it's just too early to tell.
  4. They're leaving a lot out[ Go to top ]

    Mono maybe be good stuff, but at this point it's just too early to tell.

    Having used mono heavily I can tell you right now that it works. Yes, there is still stuff missing, but the holes are getting filled incredibly fast.

    I use mono and java and .net and as usual: use the best tool for the job. (Sometimes that appears to be COBOL)
  5. Which Java?[ Go to top ]

    The "Langage" formerly known as Java, is not keeping up on the server side, loosing to PHP. There are several companies that dumed Java (ActionWatch, etc.) to go to PHP. Look at the number of applications that are easy to install in PHP (Bugs, Content, Forums, etc.) Plus PHP version 5 has SOAP built in. So server side main competion is PHP and boy they are good.Java's main advanage is JNLP, the way to web deploy applications that bypass the browser (ex JDNC), when used with SoA such as Hessian. So Java's future growth is back on the Client side, and it will continue to shrink on the server

    Is this some kind of joke? PHP killing Java on the server side? Java's future on the client using webstart?

    PHP and Java aren't really even comparable technologies. You can make some comparisons between PHP and JSP, but the Java platform as a whole provides a range of features and support that PHP isn't even targeting. PHP is great for small web based applications... like those you mentioned. But you're never going to see PHP doing the heavy lifting designated to Java at the enterprise level. And trying to co-ordinate a PHP project consisting of more than two developers is at best a nightmare.

    Java's future (and present) is on the server side. Webstart is an interesting delivery mechanism... but, you're still deploying Swing applications. Java has never done a good job on client side UI's... Microsoft products have always eaten it's lunch. I say let Microsoft have the client side... for me, all the interesting stuff is on the server side, anyway.

    //Nathan
  6. Which Java?[ Go to top ]

    Nathan,
    I hope you are not saying this based on marketing but bassed on searching for an application.
    When I client contracts me, I do a build vs buy and look what to start with, so I look at all the platforms. I won't list a lot of PHP but look at CMS:
    http://www.xoops.org

    (Compare it to.... Java's Jackrabit?)

    In general, given an application, there are more and better choices in PHP to start w/. So please pick a application area and google. I'd like to know objectively, what application area does Java have more or better production apps ?
    PHP is very very good option on server side, so if you want to be one of the late adaptor on browser apps, consider PHP to be fair to your client. They don't care about mumbo-jumbo geek stuff.

    Early adpators are bypassing the browser IMO, so Mono development ide's look interesting. (No one sees the distruptive technology comming)


    .V
  7. Java CRM?[ Go to top ]

    "In general, given an application, there are more and better choices in PHP to start with"

    It is so true. Where is the Java CRM eqvivalent to SugarCRM?

    "The project has been on SF.net since April 2004 and has had over 40,000 downloads. The SF.net team is proud to make SugarCRM the October 2004 Project of the Month."
    http://sourceforge.net/potm/potm-2004-10.php

    SugarCRM
  8. Java CRM![ Go to top ]

    --> OpenCRX
  9. http://sourceforge.net/projects/compiere/

    Smart ERP+CRM solution for Small-Medium Enterprises in the global marketplace covering all areas from customer management, supply chain and accounting. For $2-200M revenue companies looking for "brick and click" first tier functionality.

    Registered: 2001-06-08 19:37
    Project of the month for: February 2004
    License: Mozilla Public License 1.1 (MPL 1.1)

    Part of the "most active" project list on sf.net since about 3 years !!!

    Sorry, but I do realy see not reason to use PHP today !
    (Warning : bad nasty hidden troll ready to catch ya)

    :P
  10. but finished it will never be[ Go to top ]

    Part of the "most active" project list on sf.net since about 3 years !!!

    Yes and I (the fortune-teller according to some) tell you that it probably will be one of the most active in the next 3 years too!

    From one that recognizes a doomed project when he sees one. :)

    Regards
    Rolf Tollerud
  11. but finished it will never be[ Go to top ]

    <blockquote(the fortune-teller according to some)
    That would be saying that.

    So do you use a crystal ball, hold an envelop to your head, read tea leaves, or just around and yank it out (about as cleanly as I put it :) )?
  12. but finished it will never be[ Go to top ]

    Screwed that post up royally.
    (the fortune-teller according to some)

    That would be me saying that.

    So do you use a crystal ball, hold an envelope to your head, read tea leafs, or just reach around and yank it out (about as cleanly as I put it :) )?
  13. but finished it will never be[ Go to top ]

    Mark: "So do you use a crystal ball, hold an envelop to your head, read tea leaves.."

    No, it is called "statistical inference" or "interpolate a graph" or just "experience".
  14. bad joke[ Go to top ]

    Mark: "So do you use a crystal ball, hold an envelop to your head, read tea leaves.."No, it is called "statistical inference" or "interpolate a graph" or just "experience".

    wouldn't that imply having collected and verified data to generate statistics? I'm not sure what statistical inference is suppose to mean. Is that a new marketing term? Like a crazy combination of an inference engine with a statistics engine.

    </end_joke>
  15. bad joke[ Go to top ]

    Mark: "So do you use a crystal ball, hold an envelop to your head, read tea leaves.."No, it is called "statistical inference" or "interpolate a graph" or just "experience".
    wouldn't that imply having collected and verified data to generate statistics? I'm not sure what statistical inference is suppose to mean. Is that a new marketing term? Like a crazy combination of an inference engine with a statistics engine.</end_joke>
    We call it "calculated guess" down here. :)
  16. but finished it will never be[ Go to top ]

    Mark: "So do you use a crystal ball, hold an envelop to your head, read tea leaves.."No, it is called "statistical inference" or "interpolate a graph" or just "experience".

    Ok. Just asking. It was tough to tell. :) I do think "statistical inference" is much like the last one on my list though.

    Now back to the important part of this thread. Figuring out how to get O'Reilly to print more books on Java.

    Really, how many books has anyone seen on .Net that are more than VB.Net/C#, ADO.Net and ASP.Net? And on Mono for that matter? Now look at the Java section. JSF, SWT, Tapestry, Ant, AspectJ, Jini, Javaspaces, Java, J2EE, Swing, etc. If there are any, they are not at my library ... I mean bookstore.
  17. .Not and windozeINSECURE[ Go to top ]

    http://www.theregister.co.uk/security/security_report_windows_vs_linux/


    Considering the publicity that has surrounded - and, despite super new security-focused Service Packs, continues to surround - Windows security issues, Microsoft's determination to demonstrate that Linux is less secure than Windows shows a certain chutzpah. The company has however had some support here; Forrester, for example, provides some numbers that can be used to support the contention that Microsoft flaws are less severe, less numerous and fixed faster. And although there's a general readiness among users to believe that Windows is a security disaster area, there's also a reasonable amount of support for the view that Linux would get just as many security issues if it had anything like Windows' user base.

    But what's the truth? For every claim there is, somewhere, a counterclaim. But until now there has been no systematic and detailed effort to address Microsoft's major security bullet points in report form. In a new analysis published here, however, Nicholas Petreley* sets out to correct this deficit, considering the claims one at a time in detail, and providing assessments backed by hard data. Petreley concludes that Microsoft's efforts to dispel Linux 'myths' are based largely on faulty reasoning and overly narrow statistical analysis. Even if you think you know this already (as we fear may be the case for numerous Register readers), we think you'll find it useful to be able to say why you know it, what the facts and the numbers really are, and where you can get the document to back up what you're saying. Appropriately enough, we're offering the report for free. You can browse through it here, and you can download it in PDF format here.

    We encourage you all to grab a copy and give it a good read, but as a service for the fast fact junkies, we've produced a few bullet points of our own. All of these are clearly supported (unlike some similar efforts you might find elsewhere) by Nicholas' report, but don't just take our word for that, check it against the full report.
  18. Which Java?[ Go to top ]

    PHP is very very good option on server side, so if you want to be one of the late adaptor on browser apps, consider PHP to be fair to your client. They don't care about mumbo-jumbo geek stuff. Early adpators are bypassing the browser IMO, so Mono development ide's look interesting. (No one sees the distruptive technology comming).V

    Every client I've ever worked for cared about the mumbo-jumbo geek stuff.

    I remember one client rejected the idea of tacking on a web based reporting system onto a Peoplesoft app in 1999 because it was too geeky and not standard with Peoplesoft. They claimed that the users wouldn't understand clicking on a URL to get a report.

    Now I really believe they didn't want to hire anyone who knew that to support it when I was gone, so they stuck with SQR reports emailed to users in Lotus Notes as an attachment.

    Remember, this was 5 and a half years ago.
  19. Which Java?[ Go to top ]

    Yup. Java sux so you all jump on PHP. But first pass by Vic's office maybe he can sell you some consulting on Struts :-). It's cool and scalable. Hurry up, he might jump in PHP camp.
  20. Which Java?[ Go to top ]

    consulting on Struts :-). It's cool and scalable. Hurry up.

    :-) Ah... but everyone knows Struts now, so I can't bill for it.
    I am now trying to "scare up" consulting in JDNC on client side where Java is strong. Here is something you ALL can use right now:
    http://www.jext.org/apps/jext.jnlp

    On Linux, Mac and Windowze you can just get the url and start a working editor. Can't do that in Flex (or PHP). Try it next time somone sayz: "Oh, java that it the back end EJB thing"

    .V
    ps: my tone is that Sun is killing my "Langage" but I don't know how to debate "tone". just facts is ok.

    pps:
    I use PHP vBulletin (baseBeans.com) and Matins (http://neiclass.com/mantis/main_page.php) to track bugs in my JDNC application (boardVU.com). I have no idea how to program PHP becuase I can't figure out how to set up a caching DAO like iBatis.
  21. Which Java?[ Go to top ]

    Hmm. Can't we just make those archives smaller ? I mean 2.5M on a bad net links means I can't use your editor :-))

    Wasn't here a thread about some new compression format for jars, some bad ass compression that would shrink jars to ten times smaller ? I wonder if that could be used in Java Web Start
  22. Which Java?[ Go to top ]

    Hmm. Can't we just make those archives smaller ? I mean 2.5M on a bad net links means I can't use your editor :-))Wasn't here a thread about some new compression format for jars, some bad ass compression that would shrink jars to ten times smaller ? I wonder if that could be used in Java Web Start

    Yes.
  23. Which Java?[ Go to top ]

    Bingo. A data caching and compile strategy is the key that is needed for PHP to be useful. Once you have a mod_php that can access all of the Apache2 functionality, and can load your libraries as each child server is spawned, run a PHP interpreter in it's own thread, and once there is a caching server that can do lazy loading, transations, and mark dirty objects, then PHP will be a contender. As long as you can afford db queries with every request, PHP is golden. Various PHP caches are good for loading libraries, but can get ugly for complex apps like a portal, and can never be trusted for data objects.
  24. Which Java?[ Go to top ]

    I won't list a lot of PHP but look at CMS:http://www.xoops.org(Compare it to.... Java's Jackrabit?)In general, given an application, there are more and better choices in PHP to start w/. So please pick a application area and google. I'd like to know objectively, what application area does Java have more or better production apps ?

    Vic, makes a good point. There are some great PHP projects out there. XOOPS is an excellent example. phpAdsNew rules. phpClassifieds is also very nice.

    Still, if you want a list, I think one's been started with mobile applications, but here's a list of areas where I think Java excels over PHP:

    1. Desktop clients
    2. Mobile clients
    3. Embedded Graphs and Charts
    4. 3D Graphics
    5. 2D Graphics
    6. ORM
    7. Search Engines (Thank you Lucene)
    8. Mimicking vendor formats (POI)
    9. PDF Generation (iText, FOP)
    10. Cron/Daemon Processing
    11. Socket Programming
    12. Artificial Intelligence
    13. However you want to classify JSF (I would call it .NET-like postbacks, but that's not quite accurate.)
    14. Application Servers
    15. Operating Systems

    And so on. Anyone who says Java is dead is fooling themselves for reasons unknown to me.

    PHP is good for simpler websites. And the reason I say simplier websites has nothing to do with scalability. I want to see cold hard metrics on that the two languages and I fear we won't find any because PHP and Java go about their business differently. My gut tells me that a Hello World application in Java runs more slowly than a Hello World application in PHP because Java creates a lot of objects. But my gut also tells me that a large (10,000+ lines of code project) scales, is more efficient, and most importantly, is easier to maintain in Java since it self organizes better, it offers a richer API, and generally is suited to larger scale development.

    I also think that if you put a newbie's Java app vs. a veteran's PHP app, the veteran wins everytime.

    Perhaps its time for a real analysis. Is there an application which could be written and run on the same hardware to determine metrics for Java vs. PHP performance?
  25. Which Java?[ Go to top ]

    The "Langage" formerly known as Java, is not keeping up on the server side, loosing to PHP. There are several companies that dumed Java (ActionWatch, etc.) to go to PHP. Look at the number of applications that are easy to install in PHP (Bugs, Content, Forums, etc.) Plus PHP version 5 has SOAP built in. So server side main competion is PHP and boy they are good.Java's main advanage is JNLP, the way to web deploy applications that bypass the browser (ex JDNC), when used with SoA such as Hessian. So Java's future growth is back on the Client side, and it will continue to shrink on the server
    Is this some kind of joke? PHP killing Java on the server side? Java's future on the client using webstart? PHP and Java aren't really even comparable technologies. You can make some comparisons between PHP and JSP, but the Java platform as a whole provides a range of features and support that PHP isn't even targeting. PHP is great for small web based applications... like those you mentioned. But you're never going to see PHP doing the heavy lifting designated to Java at the enterprise level. And trying to co-ordinate a PHP project consisting of more than two developers is at best a nightmare.Java's future (and present) is on the server side. Webstart is an interesting delivery mechanism... but, you're still deploying Swing applications. Java has never done a good job on client side UI's... Microsoft products have always eaten it's lunch. I say let Microsoft have the client side... for me, all the interesting stuff is on the server side, anyway.//Nathan
  26. Which Java?[ Go to top ]

    Sorry for last reply with no body.

    I also can't believe how some one compare PHP to Java and than gives the verdict that it is killing Java on the server side. Even .NET haven't impacted Java a bit, and the reason being plain and simple, no body in the right mind can bet on Mono.
  27. A foot in the door is enough[ Go to top ]

    If you think Java can survive without a client, dream on. Businesses will demand better clients as per M$, so if java can't deliver a client then why would companies use java at the backend and .NET on the frontend? If you let M$ get a foot in because you can't be arsed to support a great client then they will win (web pages suck, good for internet).

    If Sun would remove their head from their arse and endorse the likes of SWT then it would suceed on the desktop as well.
  28. Welcome to the club :-)! Really is like a good joke but a tragically one; it’s so obvious that a language needs business ready and reliable API and tools equally for desktop/client-side and server-side, for surviving especially when is challenged by a competitor like Microsoft … really I’m not able to figure out what is the marvelous business strategy of SUN& Co if they assume they have a minimum chance to survive in the next 2 –3 years. Is it so difficult? I don’t think so, but how can understand anymore the business model of SUN ?! Not me …
    "To see what is in front of one's nose requires a constant struggle."
    - George Orwell
  29. you are welcome![ Go to top ]

    Cristian,

    I see that you are relatively new so I take the opportunity to say Welcome to TSS! It certainly feels good that it is at least one other intellectual here in the club! :)

    Regards
    Rolf Tollerud
  30. With clients holding state information and even off-line versions the server-side is dramatically simpler. PHP is easy to deploy anywhere. No need for a dedicated server. And scaling! I never understood this: 400 simultaneous PHP users have no problem sharing a single computer that chokes already after 40-50 with Java. But still the Java camp claims that with 1000s upon 1000s of users Java scale better. Where is the logic?

    Maybe 80-90% of enterprise applications can be build with Mono/Glade at the client and PHP at the server..
    Regards
    Rolf Tollerud
  31. enterprise applications can be build with Mono/Glade at the client and PHP at the server.

    Allready done, see C# plus PHP example app that I use:
    http://www.mantisbt.org

    .V
  32. PHP failure story[ Go to top ]

    With clients holding state information and even off-line versions the server-side is dramatically simpler. PHP is easy to deploy anywhere. No need for a dedicated server. And scaling! I never understood this: 400 simultaneous PHP users have no problem sharing a single computer that chokes already after 40-50 with Java. But still the Java camp claims that with 1000s upon 1000s of users Java scale better. Where is the logic?Maybe 80-90% of enterprise applications can be build with Mono/Glade at the client and PHP at the server..RegardsRolf Tollerud

    I don't know Rolf. I had the exact opposite experience about a year ago. A year ago, we needed a better way to distribute and manage web content on jboss.org. Instead of writing this functionality ourselves, we decided to port our website to PHP PostNuke. I'm the one who did the port.

    It took about 2 weeks to do the port to post nuke. We used the regular post nuke features plus the forum software.

    When I put the site up live to service www.jboss.org is when the problems started to happen. When we were using Tomcat/JSPs/JBoss/EJBs to service our website needs, we had about a 20% CPU utilization for our peak load of 300 concurrent users on a dual-cpu 500 Mhz machine running MySQL as well. When we switched to PHP PostNuke is when huge problems surfaced immediately. We went from 20% CPU utilization with peak 300 concurrent users to 100% CPU utilization with max ability to serve only 50 concurrent users. We quickly switched back to our old system and decided to write the software ourselves (and eat our own lunch with JBoss/Tomcat, etc...). Thus Julien Viet was hired and JBoss Nukes CMS engine was born.

    This problem of PostNuke was not that PHP didn't scale, but
    rather that the PostNuke application wasn't designed horribly. When I looked into the problem, I found that PostNuke did absolutely no caching whatsoever and was hitting the database at least 2 times per page access. The old website on Java/JBoss cached almost everything.

    Although I believe in general that PHP probably scales as well as Java performance wise, I am very unconvinced that it scales maintenance wise. A Typesafe, compiled language is so important for large scale systems. I know from experience as I've written a few decent sized Python applications. Scripting languages allow you to prototype quickly, but once there's more than one developer working on things, things can get quite hairy as you lose a lot of the structure a language like Java (or even C and C++) give you.

    Bill
  33. PHP failure story[ Go to top ]

    "Thus Julien Viet was hired and JBoss Nukes CMS engine was born"

    I guess the new system is a little larger than the old then?

    Do you still run the 300 concurrent users on a dual-cpu 500 Mhz machine with MySQL as well? With 20% CPU utilization?
  34. Java vs. PHP[ Go to top ]

    Although I believe in general that PHP probably scales as well as Java performance wise, I am very unconvinced that it scales maintenance wise. A Typesafe, compiled language is so important for large scale systems. [...] Scripting languages allow you to prototype quickly, but once there's more than one developer working on things, things can get quite hairy as you lose a lot of the structure a language like Java (or even C and C++) give you.Bill

    Since I also enjoyed some heated debates on that 'PHP vs. Java' thing on several occasions, I must say that Bill's statement simply hits the nail ...

    Regards,
    Stefan
  35. With clients holding state information and even off-line versions the server-side is dramatically simpler. PHP is easy to deploy anywhere. No need for a dedicated server. And scaling! I never understood this: 400 simultaneous PHP users have no problem sharing a single computer that chokes already after 40-50 with Java. But still the Java camp claims that with 1000s upon 1000s of users Java scale better. Where is the logic?Maybe 80-90% of enterprise applications can be build with Mono/Glade at the client and PHP at the server..RegardsRolf Tollerud

    Ever seen enterprise application written in PHP? I've developed some application in PHP. It is very useful and nice environment indeed. But it completly doesn't fit into enterprise scenario. First of all it becomes unmaintenable very fast wit 50 000 LOC it is complete nightmare to introduce some new feature, unless You and Your team *super exceptionally* careful about using PHP language.

    Yes it uses less resources and could potentially handle more requests in some cases. But becomes nightmare in terms of testability, maintenability and reilability.

    It works for some dynamic sites - even huge sites (yahoo?) but I cannot imagine who could potentially write banking application in PHP? Any examples of non-trivial PHP enterprise applications?

    Artur
  36. Ever seen enterprise application written in PHP?

    Did you see the links to CMS, CRM, Ads, Classifieds, etc in this thread in PHP? Go on google, and pick an application, you will find more and better in PHP. That is the point. I can ask you better, show me a EJB application? Please.

    I am not saying PHP is better people or Mono is better, I am saying they are very competetive. Java's one strength that I see is JDNC and JNLP, ease of deployment.

    If you write a new application, just make sure if you represent Java, that you use a effective designs (such as avoid EJB/JSF) so the people can't come after and say I wish we used something else becuase this was crap.
    And show then some RiA in Java, and they'll love you.

    And try to inform your clients that Sun != Java. I do this by using JRockit VM. This way if someone says: Sun is not working enterprise, it does not impact bad on Java.

    Java O/S tools are vastly suppeiror to ANY tools in ANY langage, but that's a bit harder sell then EJB. What a great marketing workd "Enterprise", that means a bussines app. with lots of concruent users.

    .V
  37. Vic, did you see my earlier post. You can't say that any app in Java has a better version in PHP.
  38. Hmmm[ Go to top ]

    Moderators, can you ban Rolf Tollerud from here? He pollutes so many of the discussions here that they detract from the topic under question and half teh discussions are wasted disproving here.

    Rolf, my man, you got something to say, why don't you go start a blog? I promise you, the day you stop pestering me to do so, I'll go download .NET. Heck, I'll even install Mono!
  39. Did you see the links to CMS, CRM, Ads, Classifieds, etc in this thread in PHP? Go on google, and pick an application, you will find more and better in PHP.

    I should be more precise. I talked about apps which have more than 1-2 users with more than 100-200 employees and more than 100 customers. Of course there are plenty CRM, CMS etc. apps in PHP but there is even more written in PERL. Does it mean that PERL is better than Java? I'd like to read real use success story - who use these CRMs? Any Banks? Finacial institutions?

    <blockqoute>
    That is the point. I can ask you better, show me a EJB application? Please. I am not saying PHP is better people or Mono is better, I am saying they are very competetive. Java's one strength that I see is JDNC and JNLP, ease of deployment. If you write a new application, just make sure if you represent Java, that you use a effective designs (such as avoid EJB/JSF) so the people can't come after and say I wish we used something else becuase this was crap.
    I didn't talk about any specific mechanisms in Java. Personally I'm a bit scepticall about EJB. I was talking about Java itself - in terms of reliability, supportability, testability, maintanability etc. This does matter when you write application with more than 50000 LOC. And 50000 LOC is small project. For me claiming that PHP could potentially replace Java on serverside (generally) is just funny (Hey what aboul mod_perl?). Nothing else.

    Artur
  40. fond guilty of associating[ Go to top ]

    Artur: "I was talking about Java itself - in terms of reliability, supportability, testability, maintanability etc"

    Have you ever heard the term "guilty by association?"
    It is not the Java language that there are something wrong with, it is the Java programmers!

    Java programmers are obsessed with "Computer Science", more interested in doing things the right way than getting the job done. Unfortunately the right way one year isn't the right way another year! So Java is found guilty by associating with this kind of people..

    Romain Guy,
    The Java world has in fact been built by armies of smart guys. The problem is that those people are way too smart and love clean designs way too much. They also worship Liz, (I just came up with her name so don’t even ask) the Goddess of Complexity.

    How come Java developers are so concerned by design problems? I know people who spend so much time trying to find the right architecture, the best framework, and the most appropriate patterns, and writing the test suites that they somehow forget to write the application itself. Should a few hours of work really become weeks?

    Of course any third rate psychologists can tell you were such behavior come from.

    Regards
    Rolf Tollerud
  41. fond guilty of associating[ Go to top ]

    Artur: "I was talking about Java itself - in terms of reliability, supportability, testability, maintanability etc"Have you ever heard the term "guilty by association?"It is not the Java language that there are something wrong with, it is the Java programmers!Java programmers are obsessed with "Computer Science", more interested in doing things the right way than getting the job done. Unfortunately the right way one year isn't the right way another year! So Java is found guilty by associating with this kind of people..Romain Guy,
    The Java world has in fact been built by armies of smart guys. The problem is that those people are way too smart and love clean designs way too much. They also worship Liz, (I just came up with her name so don’t even ask) the Goddess of Complexity. How come Java developers are so concerned by design problems? I know people who spend so much time trying to find the right architecture, the best framework, and the most appropriate patterns, and writing the test suites that they somehow forget to write the application itself. Should a few hours of work really become weeks?
    Of course any third rate psychologists can tell you were such behavior come from.RegardsRolf Tollerud
    Now I know why all those VB systems we have here need to be replaced one year after done! They are not built to last, according to Rolf. Indeed a good way to make money: "company business has changed, can you fix the system? No way, it is easier to throw it away and build another from scratch!". Rolf, you should know that these weeks spent trying to make the system better will pay of a 100 times in the future. But what do you know, you have the shameful job of defending the 100s of thousand VB programmers who think that ACID is a property of lemons...
  42. With all due respect Henrique, I rather trust someone like Vic, with 20 years experience and more than 100 large successful deployments under his belt.
  43. fond guilty of associating[ Go to top ]

    Now I know why all those VB systems we have here need to be replaced one year after done! They are not built to last, according to Rolf. Indeed a good way to make money: "company business has changed, can you fix the system? No way, it is easier to throw it away and build another from scratch!". Rolf, you should know that these weeks spent trying to make the system better will pay of a 100 times in the future. But what do you know, you have the shameful job of defending the 100s of thousand VB programmers who think that ACID is a property of lemons...
    I know nothing about VB systems(VB applications are short living and stanalone).I do not think web applications with EJB stuff need to be systems too. EJB is going to die (ORM is fashion at this time) so JAVA applications have the same future too. This stuff will newer replace SQL and COBOL in *systems*, forget it. Toys are designed for toys do not confuse it with systems.It is always better to replace toy with a new toy than to develop it for future, it because old toys are boring for user.
  44. fond guilty of associating[ Go to top ]

    I know nothing about VB systems(VB applications are short living and stanalone)

    You are correct. You don't. They are neither short-lived or standalone just because VB was uses.
  45. Missing so many points ...[ Go to top ]

    Rolf, 'guilt by association' is a classic illustration of a logical fallacy - an idea that seems to make sense but. when examined critically, does not. Of course you knew that, right ?

    Something that struck me the moment I entered this thread was a comment that suggested that Java would become the new COBOL ... I think the idea was that this was bad. Maybe I misunderstood ?

    In my view, there could be no more successful future for Java than to become the next COBOL. Many developers want a secure and long career, and many enterprises want rock-solid software backbones they can rely on indefinitely. COBOL has provided both of those things for much longer than most of us have been in this business. It provides them today. It would be terrific if I could work as a Java dinosaur for the next 10-20 years ;-)

    The other thing that strikes me is a bizarre bias against 'the smart guys' or 'IT engineering' on the part of many of the most virulent critics of Java on this thread. Oddly, most of these people also seem to have trouble expressing themselves in English. Perhaps these people really do dislike education ?

    There is, at present, no such thing as 'software engineering'. However it is a goal worth struggling to achieve ... there will always be a group of people who want something they can use to make money now with minimal education. There's nothing wrong with that, of course. We need both kinds of people to advance the cause of software development. But the quick-fix guys need to keep in mind that without the big brains and the - gasp ! - 'academics' - ungasp - the IT industry would not exist at all.

    As for conversations about J2EE and .NET, they all seem to miss the most important point in my view. Many very large corporations simply will never rely on Microsoft for critical systems. This is a financial and political decision not a technical decision. Microsoft's size and business tendencies make other corps in the same class wary of Redmond of course. But fundamentally it is in the interest of large-scale buyers to maintain competition between vendors, and they will do so. There's not a whole lot to think about there in my view.
  46. History lesson. Taylorism.[ Go to top ]

    One can add that the Java programmer’s way of working already got a name in 1913, when Frederick Winslow Taylor published: «The principles of Scientific Management».

    Lenin, the archetype of all well-meaning impractical theoreticians, was a big fan. The Soviet Union variant of communism is it's biggest failure to date.

    Regards
    Rolf Tollerud
  47. History lesson. Taylorism.[ Go to top ]

    One can add that the Java programmer’s way of working already got a name in 1913, when Frederick Winslow Taylor published: «The principles of Scientific Management».Lenin, the archetype of all well-meaning impractical theoreticians, was a big fan. The Soviet Union variant of communism is it's biggest failure to date.RegardsRolf Tollerud

    Theory is something that is not/can not be proven by practical experiments. Practicality is applied theory in the real world. Therefore you have to try and possibly fail many times to find a practical solution. You are feasting on this 'failure' as it was a sign of incompetence and stupidity. I think you are very ignorant and one dimensional in your assessment of things. Have you ever gone through this process yourself at work? If you have, would you call yourself an impractical theoretician?

    This is how Microsoft works. They copy proven theories from others and apply them for their own benefit. They can do this because they are procted by vast cash reserves and monopoly that is hard to shake. They have the luxury to wait for others to struggle to find practical solutions to their theories and then copy them when the fruits of that labour are ripe. Microsoft does not invent things - they copy them. The simple reason for that is that they never have to innovate to compete. They only have to be barely good enough, and the monopoly does the rest.

    Prime example of successful capitalism, would you not say that?
  48. fond guilty of associating[ Go to top ]

    Java programmers are obsessed with "Computer Science"

    Yep - and russian drinks vodka all the time, frenchman is a great lover etc. Come one - please don't use generlized quantifiers.

    After all - I always thought that there is nothing bad with "computer science" - is it?

    Artur
  49. But in this case it is the literal truth Artur, as you see the Java camp don't deny it and is even "orgulloso" of it! In spite of recent catastrophe of epic dimensions.

    Java will die, no wrong of course Java will not die but Java will loose out to .NET because of it. Then you can talk about "fallen on own behalf" or "sawing off the branch you're sitting on"!

    No other language have this characteristic, not Perl, not Ruby, not Python, C, C++,C# or any of the other .NET languages, it is a sole Java trait. And that they can not see themselves that they are the laughing stock of the industry!

    Regards
    Rolf Tollerud
  50. Which Java?[ Go to top ]

    And what about Gaming, portable devices ? Java seems to be pretty strong in these areas as well.
  51. Which Java?[ Go to top ]

    Sun's "Java" the brand?Read this: http://www.onjava.com/pub/a/onjava/2004/10/06/jds.htmlThey can't help themselfs(look at their Stock price), so they can't help others. Sun's Java in this case is a C desktop, and that is all they think of it.IBM's Eclipse JVM or BEA's JROckit VM are just fine, and GNU is cathing up. (They can't call it Java I think)The "Langage" formerly known as Java, is not keeping up on the server side, loosing to PHP. There are several companies that dumed Java (ActionWatch, etc.) to go to PHP. Look at the number of applications that are easy to install in PHP (Bugs, Content, Forums, etc.) Plus PHP version 5 has SOAP built in. So server side main competion is PHP and boy they are good.Java's main advanage is JNLP, the way to web deploy applications that bypass the browser (ex JDNC), when used with SoA such as Hessian. So Java's future growth is back on the Client side, and it will continue to shrink on the server (Unless Fujitsu KILLS Sun)Mono looks STRONG!!!!!! (JUnit, Ant, Ibatis, Spring all ported). And Mono is cross platform, Mac, Windows, Linux.Java "the brand" was killed dead by Sun's EJB marketing. KISS allways wins out. In J2SE 5.0, Java jumped the shark by not offering enough Client side, where it has an advantage.(Still time for IBM to expand things like NetX or JNLP, but.. they are stuck on their RCP, not on operations and deployment, where NetX is and the need is)How do I web deploy Mono anyone? .V

    I can't understand the tone here. Sun's business is not where it used to be, so any technology not coming from Sun is good. Mono exists, therefore it is a better choice, especially given that it decides to port existing, stable Java projects (that's simply a sensible thing to do). Serverside alternatives exist, like PHP and .NET, therefore serverside Java is doomed. EJB has problems, so J2EE is doomed whether or not .NET or other solutions provide the perfect answer that J2EE fails to.

    There's a point to not disregarding competition, but only Microsoft used to generate such antipathy based purely on religion.
  52. Which Java?[ Go to top ]

    ... Mono exists, therefore it is a better choice, ...
    Hey, I got some fantastic swamp land to sell you. Or maybe you would be interested in a bridge in NY? You seem like the kind person. Also, I have this great retirement plan. Just involves 2 bucks a week and the lottery.
  53. Nonsense![ Go to top ]

    Sun)Mono looks STRONG!!!!!!

    Perhaps you should run SciMark2 to see who really is *STRONG*. Osvaldo's brief test tells a different story. Java 5.0 server VM and MS CLR leaves the mono ameteurs in the dust. Read the Composite scores below...

    (A higher score is better)
    J2SE 5.0 (-server -Xbatch):

    Composite Score: 390.6889836966191
    FFT (1024): 197.93924031256967
    SOR (100x100): 538.5564839078068
    Monte Carlo : 67.108864
    Sparse matmult (N=1000, nz=5000): 317.8273464995424
    LU (100x100): 832.0129837631767



    Microsoft's CLR 1.1.4322:

    Composite Score: 261,40 MFlops
    FFT : 184,82 - (1024)
    SOR : 317,85 - (100x100)
    Monte Carlo : 32,29
    Sparse MatMult : 371,19 - (N=1000, nz=5000)
    LU : 400,84 - (100x100)



    Mono 1.0.2 (--optimize=aot) (**)

    Composite Score: 140,78 MFlops
    FFT : 66,78 - (1024)
    SOR : 314,55 - (100x100)
    Monte Carlo : 2,95
    Sparse MatMult : 126,32 - (N=1000, nz=5000)
    LU : 193,28 - (100x100)
  54. Which Java?[ Go to top ]

    So Java's future growth is back on the Client side, and it will continue to shrink on the server (Unless Fujitsu KILLS Sun)

    This whole comment is silly talk. I am sick of hearing Mono propaganda.

    For starters Sun and Fujistu have major cooperation agreements. They will have a shared server line for 2006.

    Their balance sheet is going to look rather different after their current x86 server sales growth (thanks to Opteron) of over 130% continues for a few more years. Things are going to get really interesting once Sun releases their "niagra" (8-cores per chip) processors in 2006, and their "rock" processor in 2008. IMHO their current issues are due to them taking so long to bring out a successor to the UltraSPARC III.
  55. Which Java?[ Go to top ]

    IBM's Eclipse JVM or BEA's JROckit VM are just fine, and GNU is cathing up.

    I ran my own tests using SciMark2 for IBM, BEA and Sun's JVM's. These were done on a 3GHz P4.
    Sun's server VM was fastest by a country mile.


    IBM JDK1.4.2

    Composite Score: 473.15337357628323
    FFT (1024): 303.3461982608695
    SOR (100x100): 446.8819307189731
    Monte Carlo : 41.695471818896095
    Sparse matmult (N=1000, nz=5000): 508.4251222223496
    LU (100x100): 1065.4181448603276

    java.vendor: IBM Corporation
    java.version: 1.4.2
    os.arch: x86
    os.name: Windows XP
    os.version: 5.1





    SUN MICROSYSTEMS: JDK5.0 -server

    Composite Score: 551.7962864593471
    FFT (1024): 311.8056013835335
    SOR (100x100): 742.8508519718736
    Monte Carlo : 16.392004458011538
    Sparse matmult (N=1000, nz=5000): 530.8707761581817
    LU (100x100): 1157.062198325136

    java.vendor: Sun Microsystems Inc.
    java.version: 1.5.0
    os.arch: x86
    os.name: Windows XP
    os.version: 5.1



    BEA Systems JRockit JDK1.4.2

    Composite Score: 372.8240492895858
    FFT (1024): 226.8783291011936
    SOR (100x100): 443.5900250731629
    Monte Carlo : 107.3741824
    Sparse matmult (N=1000, nz=5000): 448.56948350473033
    LU (100x100): 637.708226368842

    java.vendor: BEA Systems, Inc.
    java.version: 1.4.2_04
    os.arch: x86
    os.name: Windows XP
    os.version: 5.1


    SUN MICROSYSTEMS: JDK5.0 client (default)

    Composite Score: 229.93141961717885
    FFT (1024): 108.6077580589272
    SOR (100x100): 403.1286852817934
    Monte Carlo : 16.652324676588194
    Sparse matmult (N=1000, nz=5000): 158.9136732497712
    LU (100x100): 462.3546568188142

    java.vendor: Sun Microsystems Inc.
    java.version: 1.5.0
    os.arch: x86
    os.name: Windows XP
    os.version: 5.1
  56. Fac Finder #'s[ Go to top ]

    Here is a pdf on JRockit threads and gc, etc.
    http://www.bea.com/framework.jsp?CNT=whitepapers.htm&FP=/content/products/jrockit/learn/
    Things that scale,
    .V
  57. Which Java?[ Go to top ]

    Java, is not keeping up on the server side, loosing to PHP. There are several companies that dumed Java (ActionWatch, etc.) to go to PHP

    LOL

    What next...
    "Python takes over in the enterprise"
    "Visual Basic ideal for data mining"
    "Fortran taking over for GUI apps"
    "IBM, Oracle switch all their systems to Ruby"
  58. in some ways, thank god[ Go to top ]

    I'm so tired of retarded religious debates about language X over language Y. If java is becoming the new cobol, than good. That just means in 100 years, java server applications will still be doing the heavy lifting of transactional processing, just like cobol.

    GUI fads and other superficial fads are a terrible waste of time. In my biased opinion, J2EE has matured to the point where it's in Sun's best interest not to experiment too heavily with the core language or core J2EE API. That's why we have open source.

    There are plenty of developers out there doing great work. Let the community pick which one is better and once it's sufficiently mature, Sun can make it a standard. I have absolutely no problem with Sun latching onto open source work, if it makes integration easier for me in my day job.
  59. I think you need to realize one particular point here, namely the nature of the ones talking about the state of Java. O'Reilly is, of course, mostly interested in knowing in which area people will be buying books in the year to come. From that point of view the "state of Java" is perhaps not so interesting since it has had its "peak" in terms of "newness".

    In terms of selling books the things you mentioned are definitely "bad things", since *anything* new is (by default) "cooler".

    So the first question is: from what point of view do you want to discuss the "state of Java"?
  60. First of all, COBOL isn't so bad. The 'B' stand for Business so it was used as intended.

    It was possible to sense in the late 90's that the switch to Java was going to happen. Sun was very smart about shifting focus to the server side. There was a "cool" factor back then.

    So where is Java now? Is it cool?

    I think a lot of Java developers are trying to shift focus to a "lighter" J2EE stack. Many open source projects are now commonly known like Ant and Hibernate. The open source projects have become the standards IMO because they get the job done easier.

    Sun and Java have been hurt a bit by complicated high end application server solutions. Almost a mainframe kind of mentality. The current breakdown of services and components via AOP and IOC indicates current trends in software development.

    Remember, Java still has support for IBM and Oracle. The big vendor support really made Java mainstream and that hasn't gone away. The future for Java is very good. Many big companies want to run servers on commercial Unix or Linux. .Net will not meet their needs for that deployment.

    My general impression is that there is a split between the keep it simple crowd and the high end bloatware camp, mainly vendors controlling JSR for profit.

    So it's the open source projects that keep Java interesting right now.

    I'm rambling....
  61. So where is Java now?[ Go to top ]

    First of all, COBOL isn't so good, either. :-)

    I agree with your points about Java, though. As a whole, Java is doing well, and if you ask me has reached a point of maturity. There's still a lot of cool stuff being done with Java, a lot of bleeding-edge things. But the core of Java has really solidified.

    JDK 1.5 (what it should be called) is an awesome release, the EJB situation is starting to rectify itself, the open source community (not the radical zealots, but the everyday back-scratchers) has embraced it with libraries, tools, and R&D projects. And the vendors are supporting it well and in many cases innovatively, though I agree bloat and selfish interest can come into play there, but the point is they don't control Java and the open-source community has really forced them to change the way they do things.

    Sun hasn't benefitted as from Java as much as "we" have, but despite the complaints of some they've really done well in keeping the ball rolling and ensuring Java's future. I really think that Java is coming into a new growth phase that should see some exciting new stuff created, though with anything you can still expect a few bumps in the road. But Java's not going to get derailed, and certainly not boring or "dead" as some claim.

    Now regarding O'Reilly, they just need to get some creative new Java books out, not just 18th edition of such-and-such API. They're the ones getting boring, if they're not careful. It would be a mistake for them to ignore Java.
  62. Aaah, people keep talking about MONO like it is the next messiah. Well, it seems like everything new is cool...but hey, that's how they make TV too, just spin off from the original show...except Mono is a spin-off from a spin-off - all hype and fluff the whole thing.

    There is nothing cool about Mono itself but where it runs... but wait a minute, there is also this thing that starts with J, which has done the same thing a long time ago...I guess people forgot.
  63. Aaah, people keep talking about MONO like it is the next messiah....except Mono is a spin-off from a spin-off - all hype and fluff the whole thing..
    All hype and fluff? You gotta be kidding!
  64. A global invasion[ Go to top ]

    I determine a revolution to be over when it has been overthrown. So in this regard I have one question for the forum:

    What language allows you to write GUI applications, applications for mobile devices, Client-Server applications, Web applications, Web Services, and server side enterprise components regardless of operating system?

    Your answers will determine the validity of the debate.

    John Murray
    Sobetech
  65. Nice topic, it remembers me that couple of days ago I posted on the tread “Ask TheServerSide: What topics would you like to see in articles” that the moment for Java client side has come, even if really Swing has enough problems, there are many job postings requiring Swing and NetBeans Platform SDK(Core and IDE API) in North America and Europe. In the same time is possible to find a lot of job postings requiring .NET experience accompanied or not by MCAD/MCSD certifications, in North America and Europe.
    Taking a look at the immediate reality is easy to see next facts:
    - The hardware becomes fast commoditized; the multi-core chips and fast memory will bring to the desktop the power of the servers. It is a business reality with immediate effect that the traditional style of writing Desktop application or Thick Clients will switch from the usage of the OS API C/C++ based to the much more productive and secure way of memory managed environments like Java and .NET, Why? Because this creates a really huge new business opportunity and naturally, people want to make money from each new market. In reality the battle will be not so much at the level of OS’s ( Windows Server 2003 versus Linux ) only because really there are equal business advantages and disadvantages in both sides, and the deployment of managed environments is platform independent! In server-side J2EE versus .NET, will continue to coexist with a tendency of increase for .NET because people will get more and more experience using .NET and for managers will be more and more easy to decide to use .NET when really .NET makes the money, and the solution will not be anymore decided to be J2EE only because J2EE is a “business standard”, remember that always money counts and when J2EE makes the money :-) the solution will be J2EE, that’s the natural selection, nothing else, the market will not have a clear winner, just proper solutions. The real battle will be JVM+Java versus CLR+C# in the desktop/client side, because of a virtual unlimited number of existing applications requiring to switch to managed environments or completely new applications. Reasons are a lot, but security is the most important one, hard-times will come for Antivirus vendors :-). Will make no sense to not implement or in Java or in C#, even multimedia applications and games, and here the end-user-buyer is the king, if an application gets fast to market and the users are satisfied that’s all … the development time, the functionality and reliability of the end-products will create the measure between Java and C#. Of course that Windows Forms is not a perfect API, is not perfect functional, but being created, supported and improved by Microsoft, the king of desktop API, in this moment is really challenging Desktop Java only because SUN & Co. seems that they loosed their heads and they can’t understand that without improving Java on the client, here the winner and still undefeated champion will remain Microsoft, which allocates continuously huge resources to the desktop improvement. Instead complicating not necessary the language and inventing new server-side pseudo-technologies SUN &Co should concentrate and allocate resources for improving the client-JVM performance, the native Java LAF and Swing (or a new/replacement API for Swing), and desktop development tools able to compete successfully with Visual Studio .NET.
    - SWT … I like Eclipse, I use for my Java development each day but, really for a dummy user, what will make really sense? To distribute a runtime JVM with included support for desktop applications, or to complicate the instalment with the distribution of the binaries for SWT ? No natural sense, learn from Microsoft! KISS …
    - PHP, nice, and has his place and business, but let’s be honest will never be a real competition for J2EE or .NET, will be used only where will offer the proper solution, and never in big business areas like banking, insurance, booking systems etc. this systems need first very reliable storage and if you afford to use Oracle, DB2 or MSSQL Server, you go for PHP?! Have fun :-)
     So, I think that the desktop development support will be the decisive criteria concerning who will win the battle, Java or C# … but if Java looses this battle dramatically, KO, will be in danger to loose by KO the server-side too, because remember, C# runs in a managed environment, and time is going and things are improving, and Microsoft has never lost a war, but only small battles :-)
    Respect and regards,
    Cristian
  66. deja vu[ Go to top ]

    "inventing new server-side pseudo-technologies"

    I wish I'd said that!
    You will Oscar, you will..
  67. Firmly off-topic[ Go to top ]

    Rolf, you shine out like a shaft of gold when all around is dark!

    'Oscar Wilde' Run Closes After One Night

    best,
    assmund
  68. Firmly off-topic[ Go to top ]

    Rolf, you shine out like a shaft of gold when all around is dark!'Oscar Wilde' Run Closes After One Nightbest,assmund

    All that glitters isn't gold. And without light, nothing glitters or shines.
  69. Even more firmly off-topic[ Go to top ]

    Rolf, you shine out like a shaft of gold when all around is dark!'Oscar Wilde' Run Closes After One Nightbest,assmund
    All that glitters isn't gold. And without light, nothing glitters or shines.

    Mark, I guess you didn't get Rolf's joke... but it was gold to me.

    best,
    assmund
  70. Even more firmly off-topic[ Go to top ]

    Rolf, you shine out like a shaft of gold when all around is dark!'Oscar Wilde' Run Closes After One Nightbest,assmund
    All that glitters isn't gold. And without light, nothing glitters or shines.
    Mark, I guess you didn't get Rolf's joke... but it was gold to me.best,assmund

    Guess you don't know what Fool's gold is or how light reflection works.

    Glad you enjoy Rolf's "intelligent" wit.
  71. Incredibly off-topic[ Go to top ]

    Rolf, you shine out like a shaft of gold when all around is dark!'Oscar Wilde' Run Closes After One Nightbest,assmund
    All that glitters isn't gold. And without light, nothing glitters or shines.
    Mark, I guess you didn't get Rolf's joke... but it was gold to me.best,assmund
    Guess you don't know what Fool's gold is or how light reflection works.Glad you enjoy Rolf's "intelligent" wit.
    Guess you don't know Monty Python when you see it. Or, more likely, you just don't have a sense of humor.
  72. Apology[ Go to top ]

    With a few minutes to think about my last post, I regret trying to engage in a name-calling match in this forum. My apologies to you, Mr. Nuttall, and to the TSS community.

    best,
    assmund
  73. Apology[ Go to top ]

    With a few minutes to think about my last post, I regret trying to engage in a name-calling match in this forum. My apologies to you, Mr. Nuttall, and to the TSS community.best,assmund

    That's ok. It was fun while it lasted. :)

    BTW, love monty python and do have a sense of humor (See most of my posts). And very odd one at that. I like Steve Martin too.
  74. the deployment of managed environments is platform independent!

    It isn't - not with .Net.
    To distribute a runtime JVM with included support for desktop applications, or to complicate the instalment with the distribution of the binaries for SWT ?

    You have to distribute a runtime with .Net too. It's only shipped with the newest machines, and on the majority of PCs it would have to be downloaded, just like a JVM. Of course, .Net is of no use at all on the increasing numbers of Linux desktops and Macs.
    Microsoft has never lost a war, but only small battles :-)Respect and regards,Cristian

    Microsoft has lost many battles! I remember how, ten years ago, Windows NT was going to 'Kill UNIX'!
  75. the deployment of managed environments is platform independent!
    It isn't - not with .Net.
    It can be done, see Mono, even if not complete for desktop is enough ...,
    so why not, in the future, a .NET for Solaris, AIX, HP-Unix, etc. ... it is possible ...
    To distribute a runtime JVM with included support for desktop applications,
    or to complicate the instalment with the distribution of the binaries for SWT ?
    You have to distribute a runtime with .Net too. It's only shipped with the
    newest machines, and on the majority of PCs it would have to be downloaded, just like a JVM.
    Of course, .Net is of no use at all on the increasing numbers of Linux desktops and Macs.

    On the old machines, yes ... but I made reference to the next 2 or 3 years when manny of this old maschines
    will be replaced with much more powerful ones ... with Microsoft .NET CLR pre-installed. If you wait
    for the Linux Desktop or Mac OSX to replace Windows on more than 50% of user desktops, I'm sure this
    is a nice dream for the next 6 years minimum :-)
    Microsoft has never lost a war, but only small battles :-)Respect and regards,Cristian
    Microsoft has lost many battles! I remember how, ten years ago, Windows NT was going to 'Kill UNIX'!
    Pls, don't confuse marketing with reality, MS has 90% of the desktop market, so MS can control and protect
    his market part very easy, if you don't offer a general equal replacement, no change will be, users are
    tired of marketing, they want results/cheap/productive. Don't get me wrong, I'm a BEA certified ingineer,
    and I do Java for more than 5 years, I just speak from my business experience, and my view of the market, that's all ... not speaking abouth the realities is infinite dangerous in business, is fatal, and in this moment Java can loose the battle to C# only because SUN & CO. pays no real attention to desktop as business, what we have on the server-side now is enough for many many years, is the moment to start to drive the eforts to desktop! Why I say that ? The times we live now are changing the way the business is driven - see Harward study concerning disruptive technologies and economy - the business must be very flexible, short term planned, low cost and 100% success oriented, I know that sounds very managerial, but this are the times we live, is our reality each day, and we must be very flexible if we want to keep our jobs and we want to have a personal life in the same time. We started to live in a new economy and, as Rome, the old economy disapeared, what was before 2000, will never come back. I stated that the hardware evolution will generate in my oppinion the next IT boom, I think somewhere in the next 2-3 years, the desktop/thick client application business will be a great business opportunity, but the hard work will go in your hands as ingineer/developer, for a manager will be easy to decide, but for you ? What would you preffer, as a respectable ingineer who can't not afford to lie abouth his professional knowledge, to spend your nights trying to keep-up in the same time with server-side and client-side Java and server-side and client-side C#, will you not go crazy when you see that a J2EE module + Java client developmet takes you 3 times more than a COM+ .NET + C# Client module ? Remember that you must be flexible, and no manager will accept that you can't be because you decided to use a tehnology which doesn't support his business requirements ... that's it, ... its about employability and human factor, the most decisive one in business, if you are a normal person, with a family, you can't afford to do the effort to be equal productive for Java and C# if this means to sacrify your life, here a normal person will have a look at balancing between C#, Java and his life ... simple, is more than a person can do naturally, no one can handle professionally both server-side and client-side C# and Java ... is not realistic ... and the natural selection will dictate for the one who offers the best support for the requirements and for your development needs, your flexibility and employability value will be dictated by the level of knowledge in server-side and client-side, if you don't learn both sides you are in danger to search in a high pace for different jobs for different companies but only in the area where you are employable, for example server-side ... is this what a normal person expects for his future? Not me for sure ... a programming language commes with API, tehnologies, development tools, with plus or minus in all of them, but they must be very well known by a professional, and again this is natural impossible to know in the same time all Java & Co and C# & Co ... just my 2cents ...
    Respect and regards
    Cristian
  76. the rarest of the rare[ Go to top ]

    Malinescu!

    "not speaking abouth the realities is infinite dangerous in business, is fatal"

    you are actually intelligent!
  77. I have been working with Java for 8 years. And 6 months ago I decided to build an Intranet P2P Backup application for Desktops and Laptops. And I wanted my application to run continuously in the background without being intrusive to the user. It has to have a very small foot print etc. etc.

    And needless to say I had to go back to C/C++.

    I am well aware of the productivity benefits of Java. But it drinks memory for being a Desktop application, slowing down every other app. No wonder Google's Desktop Search is not written in Java.

    So pretty much in my opinion, Java will die in the Client side be it Swing or other apps working in the background.

    And on the Server side, I am not so sure. It is hard to argue about the huge productivity benefits of using Java. But I think the productvity benefit comes because of its extensive APIs and its automatic Garbage collection. But when it comes to Database centric apps, Java does not have a good solution. EJB, JDO etc. are all patch work. Unless there is some revolutinary data driven paradigm here, I can not help think that something else will marginalize Java in a few years time.

    Sekar.
  78. And needless to say I had to go back to C/C++.I am well aware of the productivity benefits of Java. But it drinks memory for being a Desktop application, slowing down every other app. No wonder Google's Desktop Search is not written in Java.So pretty much in my opinion, Java will die in the Client side be it Swing or other apps working in the background.

    Don't forget about three things:

    a) swing is getting faster and faster

    b) concept of shared classes (region in memory shared by multiple instance of jvm's) is going to be improved

    c)( this is the most important) If you are going to buy new computer today it is going to have 512 MB - 1024 MB of RAM. In 2 years it will be probably 2 GB or more. Memory footprint of 20-30 MB will be really nothing. And 30-30 is already quite a lot for small Swing application.

    Michal
  79. java/swing is getting faster and faster? I have been hearing this line for the past 8 years. I guess I have to wait 8 more years to really see java/swing get faster. In the meantime, I am using C#/.Net.

    -M
  80. java/swing is getting faster and faster? I have been hearing this line for the past 8 years. I guess I have to wait 8 more years to really see java/swing get faster. In the meantime, I am using C#/.Net.-M

    I think you missed my point.

    What i said is that it is really irrlevant how fast swing is and will be
    With the newer hardware you will be abel to develop fast enough GUI solutions using the execution model used for java 1.0 (interpretation of the code). The same applies to memory consumption. When you computer has 10 GB of RAM you don't care that much if some program takes 20 MB or 50 MB of memory.

    And coming back to Swing have you tried to compare Visual .NOT with for example IntelliJ IDEA? Can you see some substanial differences in speed?

    Michal
  81. Cristian,
      See Peter's post on the proper use of paragraphs in large posts. I had to slap myself on the back of the head a couple of times while attempting to read your posts to a.) Uncross my eyes and b.) wake up.

    Mark
  82. Cristian,&nbsp;&nbsp;See Peter's post on the proper use of paragraphs in large posts. I had to slap myself on the back of the head a couple of times while attempting to read your posts to a.) Uncross my eyes and b.) wake up.Mark
    Can you keep on topic? I will appreciate that, really I'm not interested if you sleep or not ... :-)
  83. Can you keep on topic? I will appreciate that, really I'm not interested if you sleep or not ... :-)

    Well, at least you could tell that I wasn't on topic. ;)
  84. Can you keep on topic? I will appreciate that, really I'm not interested if you sleep or not ... :-)
    Well, at least you could tell that I wasn't on topic. ;)
    Well, at least I can tell that you continue to be out of topic ... no time from my side to loose in this way ... as far as I remeber this is a discussion about IT engineering and not football bashing ...
  85. IT engineering ...
    Nope. It was about the State of Java. And from what I can tell it was more about having fun (doing cool stuff) more than IT engineering. And that is what football is for, is it not?

    Besides, I don't see how, nicely/jokingly, telling you that your post was difficult to read/follow was really off point.
    If it was important enough for you to post it, don't you think it would be nice for others to be able to read it?

    Next time I'll post it 2 ways. Once for those with the ability to understand humor and once for those who don't.
  86. IT engineering ...
    Nope. It was about the State of Java. And from what I can tell it was more about having fun (doing cool stuff) more than IT engineering. And that is what football is for, is it not?Besides, I don't see how, nicely/jokingly, telling you that your post was difficult to read/follow was really off point. If it was important enough for you to post it, don't you think it would be nice for others to be able to read it? Next time I'll post it 2 ways. Once for those with the ability to understand humor and once for those who don't.
    No problem, but I think the topic is so serious that is quite hard to think first about formatting, then secondary about the content and after about the funny factor :-), let's go back to the topic ... really this is about how our life will be affected by the next pair of years and the platform evolution ... Friends, OK :-)? Maybe a virtual-beer :-) would help to interact in this virtual-discussions but we must wait for the technology evolution and beer interface creation ... maybe a new SUN server-side technology :-) "BAO - Beer Access Objects " ?
  87. Is the Java revolution over? Has the big business-friendly approach turned Java into the new COBOL? Is the momentum behind .NET and Mono stealing away the cool factor that Java once enjoyed? Are those even bad things? What do you think?

    What an odd thing to suggest after the most significant release of Java (1.5.0) in years. There may be momentum behind .Net and Mono, but they are of little relevance to Java because they are being used for different situations. I would be very interested to see large-scale .Net enterprise deployment on Linux, Solaris or Aix. Well, of course that is impossible, as .Net is Windows-only. .Net is doing a good job of replacing Visual Basic, but that is a market which Java hardly entered. As for Mono, its a great technical achievement, but its also a scaled-down version of .Net that is developed without the approval of the producers of .Net, so is always playing catch-up and always under the threat of incompatibility, so bears little comparison with Java, where portability is encouraged and supported by Sun.

    Java is expanding rapidly into new areas: mobile technologies (Java Phone applications for example), real-time development, numerical work, scientific and engineering. New frameworks and open-source projects are starting up all the time - look at the success of Spring, for example. Look at the range of new scripting systems for Java and languages that run on the JVM and integrate with Java.

    The COBOL parallel is very strange - COBOL is a language with a very limited range of application: financial and data processing. It cannot be sensibly compared with Java - a general-purpose language with modern and evolving features. Does anyone use COBOL for real-time embedded applications? For compiler writing? For graphics? Multimedia applications?
  88. .Net is doing a good job of replacing Visual Basic

    And really very slow at that. There are still many VB and ASP apps out there. Many companies are slow (and one say "dead stop") to move to .Net from older MS technologies.
  89. The thing is, that COBOL had it right. Business applications manipulate fields of text and numbers. They don't create object heirarchies. They perform transactions on on data. How you display your data is dependent on the context, but how you manipulate it depends on your business rules. The two are not the same. And java tries to force them together.
  90. "Computer Science" in business apps?[ Go to top ]

    The OO paradigm shines in some context. For instance graphical software - object, please draw yourself! But applying OO to business application is stretching it. To compare ridiculous "Domain Objects", simple classes with a myriads of getter & setters with the elegant solutions for Auto-Cad programs! And what do you ever have for use of being able to inherit a form?

    The whole shebang reminds me of pitiful attempts of the so called "soft sciences", psychology, sociology,etc, to ape after the real sciences physics, chemistry and mathematics in a longing to be called science.

    Regards
    Rolf Tollerud
  91. The OO paradigm shines in some context. For instance graphical software - object, please draw yourself!
    no, it does not shine in this context too. MVC was invented to solve it. Model is just a model, it does not need to be OO, it just need to be valid.
  92. It is absolutely unbelievingly difficult for me to understand what that it is possible to get away with and call it "trying to shift focus to a "lighter" J2EE stack" or that "the EJB situation is starting to rectify itself", when we are talking about one of the most flagrant cases of conning ever done in history, equal to the tulip craze in Holland in the 17th century.

    And that the people responsible for "inventing new server-side pseudo-technologies" and wasting hundreds of billions of dollar for the customers, keep on working as usual, write post in forums and go along with their everyday chores, is for me totally unacceptable.

    The only thing I can think of that is comparable is the collapse of the Soviet-Union. Even in that case many western leftist intellectuals morally responsible (at least) for the atrocities and suffering of the people also go around, still publishing articles as nothing has happened!

    Tar and feather should be their rightful destiny as should be the vendors and proponents of EJB!

    Regards
    Rolf Tollerud
    (indignant citizen)
  93. P.S.[ Go to top ]

    How can you expect Java to come unscathed out of such a catastrophe?
  94. ...one of the most flagrant cases of conning ever done in history, equal to the tulip craze in Holland in the 17th century ... wasting hundreds of billions of dollar for the customers...comparable is the collapse of the Soviet-Union.

    I think that comparing the effects of EJBs to the collapse of the Soviet Union is perhaps just a little over the top? I just can't imagine reading the headline "Government overthrown due to innapropriate use of Java enterprise technology".
  95. Please read what it says.

    I do not compare "EJB with the Soviet Union". I compare the behavior of the earlier proselytes and advocates: to act as nothing has happened.

    I would have expected at least an excuse! :)
  96. Just a tiny whiny excuse for waisting approximately 300 billion dollar!
  97. I think that comparing the effects of EJBs to the collapse of the Soviet Union is perhaps just a little over the top?
    Not for Rolf. Pretty lame comparitively. :) At least it was a real event and not some fairy tale.
  98. It is absolutely unbelievingly difficult for me to understand what that it is possible to get away with and call it "trying to shift focus to a "lighter" J2EE stack" or that "the EJB situation is starting to rectify itself", when we are talking about one of the most flagrant cases of conning ever done in history, equal to the tulip craze in Holland in the 17th century. And that the people responsible for "inventing new server-side pseudo-technologies" and wasting hundreds of billions of dollar for the customers, keep on working as usual, write post in forums and go along with their everyday chores, is for me totally unacceptable.The only thing I can think of that is comparable is the collapse of the Soviet-Union. Even in that case many western leftist intellectuals morally responsible (at least) for the atrocities and suffering of the people also go around, still publishing articles as nothing has happened! Tar and feather should be their rightful destiny as should be the vendors and proponents of EJB!RegardsRolf Tollerud(indignant citizen)

    I'm gonna go out on a limb here and say you totally miss the point of EJB. You wouldn't be the first person. The first time I used EJB on a project all of us made the usual mistakes. Our team learned the hard way.

    My understanding of transactional processing and how EJB came to be doesn't fit your definition. I don't know if you've ever used EJB's. If someone gives me the task of scaling simple transactions to a single database with a clean and simple model, the obvious choice is a simple transaction monitor like tuxedo or COM+.

    what would inappropriate, would be managing complex integration and trasactional requirements, where there's multiple databases and conflicting database models with just a transaction monitor.

    Say you have 10 database and all of them have different database models. In a typical banking situation, all transactions have to be logged for auditing purposes. Lets say there's 5 separate systems a single transaction has to integrate with. all of them have their own transaction log. Now, say an update/insert to one of the systems fails. In the scenarios I am familiar with, the failure must be logged in the transaction log, but not all inserts must be rollback. I won't bother enumerating the complexies, it's rather ugly.

    for example, all transactions log entries must be maintained. In this kind of situation, a single action by the user involves over a dozen insert/updates. Now obviously, one could write a heavy client to manage all this. It would be fairly easy to scale, but the downside every customer of the bank would have to install a heavy client to use electronic banking. what is the cost of upgrading the software and how should the bank provide support for a heavy client?

    now lets say the management declares, "the business logic of managing complex transaction must be in a middle tier" how are you going to solve the problem? Would you do it in COM+? From all the resources I can find and from the COM+ API that I've looked at, it would be difficult. My bias perspective is that it would be mis-using COM+.

    could you do it with tuxedo? I would say it's probably mis-using tuxedo. What options are left? I could write my own container to manage complex integration requirements, or I can see if EJB's fit the needs. If I try to write my own container, it will take me years. If I use an existing EJB container, it saves me a ton of time.

    having said all that, EJB are good for a specific type of problem. Too many people, including myself mis-used EJB in the early days. Is that the fault of the vendors? Honestly, that's hard to say. Vendors have to sell and make money and often programmers get sucked in. that's standard business practice and over selling is part of the game.

    a lot of people like to bitch about how hard and bloated EJB's are. Well, these kinds of problems are hard no matter what platform you use. I don't consider myself a fan of EJB, but I atleast take time to understand what kinds of problems it solves. Without spending time to really understand the motivation and reason for EJB, it's easy to think all transactional processing is simple. I know of a couple financial institutions that rely on EJB's exactly for this type of problems. I've also heard horror stories of people using EJB to handle simple web trasactions, when they don't need it. I happen to focus on difficult and challenging transactional processes, so I see a place for EJB. Before I apply EJB's I make sure the need is real. As much as you complain about mainframes, ejb and cobol, most transaction you make on a daily basis eventually goes through a mainframe. Without it, many of the things you take for granted would not be possible.
  99. three famous words[ Go to top ]

    Peter give up.
    Why don't you learn from politic?

    Everytime a politician is caught with something all they have to say is that "Ok I'm guilty! - The temptation was to strong" and they are all forgiven. It is human to err and all that.

    But those that persist denying are always dismissed, for a total catastrophe. (Nixon etc etc).

    Java has improved much since 1.3 and now there is Spring and all those modern frameworks. But if you persist in defending EJB and in extension the Big Elephant Servers, you (together with the other die-hards like you) will drag Java down to the extermination or even worse, a academic teaching language like Pascal once was. :)

    You just have to be able to say the three famous words:
    "I was wrong"

    Regards
    Rolf Tollerud
    (we will forgive you)
  100. three famous last words[ Go to top ]

    Heh, some people are not just confusing Entity Beans with EJB or EJB with J2EE, now they are mixing EJB with Java the language, community and tools as a whole. Some may be still confusing it for an island even!! :)

    Whether EJB ever dies, it won't mean J2EE will also die, even less Java as a whole. There are so many other fronts where Java is advancing (cell phones, telecom systems, etc.), it is BS to say Java can't live without EJB. Quit that nonsense, Java is not just about EJB, or even just the server side.

    To those saying PHP is the way to go, don't forget that Java is not just the language: add the tools, the frameworks, the community, JSRs (is there anything like OSS/J in PHP, for example?), industry, etc., and compare all that to what PHP has today.

    For Java to die, every (I mean any and all) tool, framework, existing system, project, community and company involved with it should go under before that. How likely is it?

    Regards,
    Henrique Steckelberg
  101. three famous words[ Go to top ]

    Peter give up. It is human to err and all that.But those that persist denying are always dismissed, for a total catastrophe. (Nixon etc etc).Java has improved much since 1.3 and now there is Spring and all those modern frameworks. But if you persist in defending EJB and in extension the Big Elephant Servers, you (together with the other die-hards like you) will drag Java down to the extermination or even worse, a academic teaching language like Pascal once was. :)You just have to be able to say the three famous words:"I was wrong" RegardsRolf Tollerud(we will forgive you)

    that is a pretty pathetic response. you can do better than that Rolf. since you appear to be such an expert in financial transactions, please explain how you would achieve the scenario I described in my previous post with spring, php, C++, COM+ or any technology of your choice.

    I gladly admit I've been wrong in the past, but I also have no problem accepting that some problems just are hard to solve. No matter how hard I try to stay with KISS philosophy, the kinds of financial transactions I've been working on are complex. Could it be done with Spring? Honestly I don't know. I'm sure if I spend enough time with Spring I can wedge it in somehow.

    I'll be even more specific with the scenario. How would you solve this problem.

    1. 5 different systems. some are not in your control
    2. each transaction requires multiple insert/updates to each system
    3. a transaction must not be committed, if the transaction log fails. ie, an insert to the transaction log database fails. therefore, if 3 inserts are performed with the same connection, you can't just rollback all three.
    4. some database tables have triggers, which rely on data in other tables. therefore, some inserts are dependent on other inserts and sequence is critical
    5. the transaction has to finish within 30 seconds
    6. the systems are remote
    7. the commit threshold depends on the context of each transaction, and therefore varies significantly.

    Now let's say the user wants to order new checks. Sounds simple right. It isn't uncommon for the backend process to go through these steps

    1. the database for the account management website has to be updated (system web). usually the database is not the accounting database.
    2. an entry must be made in the electronic fund transfer database (system eft). 1 for withdrawing money from customer account, 1 for transfer to checking processing company, 1 for service charge to the bank
    3. an entry must be made in the accounting database (system M)
    4. an entry must be made in the transaction validation database (system DW)
    5. an entry must be made in the data warehousing database
    6. a call to the check processing system (system CP) must be made. if the system is down, the transaction has to be saved and processed in a batch later.
    7. a entry must be made for the customers account info in system CP
    8. an entry must be made for a new order in system CP
    9. the order number returned by system CP must be saved to system Web, EFT and M.

    that all seems easy right. by the way, this is a simple one. there are far more complex scenarios. I know of other integration scenarios, which are even more complex in telecomm's. Go ask sprint, att, mci and verizon how easy it is to integrate 2 dozen systems for something as simple as getting a new phone line.

    I would love to find an easier way to do these types of integrations. for the record, I usually tell people to avoid EJB and make them prove to me the need for EJB is real.

    if the project doesn't need it, I usually run benchmarks and offer clear explanation of how to achieve the requirements without using EJB. On the other hand, if the need is real, complexity can't be avoided. It's silly and foolish to think using a light-weight container makes solving the problem easier. It changes the details and how to implement a solution, but really it's just swapping one set of problems for another set of problems. I happily criticize and defend EJB, but atleast I try to back it up with details and not random gibberish.
  102. managing complex integration and trasactional requirements, where there's multiple databases and conflicting database models with just a transaction monitor.Say you have 10 database and all of them have different database models.
    ....
    now lets say the management declares, "the business logic of managing complex transaction must be in a middle tier" how are you going to solve the problem?
    ....
    If I use an existing EJB container, it saves me a ton of time.
    ...
    I don't consider myself a fan of EJB

    So if one has many databases then they can use outside transaction manager?
    A: 99% of J2EE applications are against a SINGLE DB, where the DB has commit/rollback, so it's silly to use Java Transactions.
    B: If in theory, some benefit was in doing trnasctions in Java... compare EJB to iBatis transactions (or Spring), which would you like then?

    .V

    ps: Who is FactFinder, CM? FactFinder, those are tecnicalities. What matters is the application not systems. Tell me about EJB applications, like we have PHP applications. Bits and bytes is for vendors. So where is an EJB application that I can download and install for a business use, since "the Langage formarly known as Java" has productivity. Goodbye Sun. Ist community and IBM ask for Java to be open source, but now its a bit more no? The Sun application server, how many production sites relative to another vendor? And JSF for "Rich UI"? Yes, do a muti row update w/ EJB and JSF.
    When a server runs, interger calcs are itrelevent. What matters is things like thread pooling and GC, and I think we all know that well.
    And here is more on JVM's:
    http://linuxintegrators.com/blog/acoliver/code/?permalink=0178.html
  103. A: 99% of J2EE applications are against a SINGLE DB, where the DB has commit/rollback, so it's silly to use Java Transactions.B: If in theory, some benefit was in doing trnasctions in Java... compare EJB to iBatis transactions (or Spring), which would you like then?.Vps: Who is FactFinder, CM? FactFinder, those are tecnicalities. What matters is the application not systems. Tell me about EJB applications, like we have PHP applications. Bits and bytes is for vendors. So where is an EJB application that I can download and install for a business use, since "the Langage formarly known as Java" has productivity. Goodbye Sun.

    like I said. if the transaction requirements are that simple, one would be crazy to go down the EJB route. If someone told me to go that route I would do my best to convince them not to. the important thing here is to have data and benchmarks to back it up. In situations where I see technology being mis-used, I try to take the requirements and run some benchmarks simulating the scenarios.

    this way, I can make a relatively well balanced presentation with solid numbers, clear explanations and the impact of choosing a simple design vs an EJB design. If the management refuses to listen, then it is out of my hands. for those 1% (which I don't believe is accurate), would you rather build your own container?
  104. for those 1% (which I don't believe is accurate), would you rather build your own container?

    Of course not. But ... It's a good question: would I use Java for it or something else ? If Java, then a DAO or ... I think the answer is JTA maybe but it would depend on a situation. I don't think I would use EJB for any kind of transactions.

    And for remote access, SoA (such as Hessain) serves me fine, so I consider myself a smart person, but I can't figure out the business use case for EJB.

    Peter, I worked in consulting for 20 years and never wrote a distributed db update, but in school we talked about it. Did you ever write a production app with muti db update or a variant?


    EJB < .NET < PHP < Java O/S


    .V
  105. yup, on a couple different occasions[ Go to top ]

    Of course not. But ... It's a good question: would I use Java for it or something else ? If Java, then a DAO or ... I think the answer is JTA maybe but it would depend on a situation. I don't think I would use EJB for any kind of transactions.And for remote access, SoA (such as Hessain) serves me fine, so I consider myself a smart person, but I can't figure out the business use case for EJB.Peter, I worked in consulting for 20 years and never wrote a distributed db update, but in school we talked about it. Did you ever write a production app with muti db update or a variant?EJB < .NET < PHP < Java O/S .V

    back when I worked on a wireless portal, we did have to integrate with infospace, the service provider's accounting system and other third party systems like booking travel. doing this kind of stuff in semi-realtime is a real pain. making sure it's reliable and scalable is also a real bitch. Something as simple as sharing personalization data across two separate systems was a source of a lot of headaches.

    the last 2 years I've been working on financial compliance, so distributed transactions and multiple database models is the norm and not an exception. I wish to god it wasn't, but no amount of wishing is going to change it.
  106. Rolf: And that the people responsible for "inventing new server-side pseudo-technologies" and wasting hundreds of billions of dollar for the customers, keep on working as usual, write post in forums and go along with their everyday chores, is for me totally unacceptable.

    Even though you have zero concrete evidence that Java or any Java related product is soley responsible for billions of dollars lost in the IT industry, I'll play along just for fun.

    Do you also find it unacceptable that over the years, Microsoft based technologies have cost businesses hundreds of billions of dollars over the many years of script kiddies writing viruses that literally take major companies offline/down for hours, even days at a time?

    All the while, Microsoft continues on, as usual, to release environments that easily allow developers/hackers/script kiddies to access kernel/system level functions and in effect side-step the whole point of developing *within* a managed space. I always thought the point of a CLR/VM was to abstract the system from the run-time. What then is the point of an abstraction if it can just be ignored at a developers whim?

    There are numerous reasons why managed environments are preferrable. These reasons are exactly why Microsoft copied the idea in the first place. The fact that Microsoft makes it easy to call system DLLs from their VM is a strong indicator to me that they still don't get it.
  107. Obviously you can not say what the state of Java is without comparing it to its competition.

    People are assuming that .NET will succeed on the desktop. Yes, it will succeed, but it will be a rewrite of old applications written in C/C++ with new C/C++ applications. Will you see C# applications? Is everythign going to be C#?

    My bet is that you will not see much of C# on the desktop at all, and this is why I believe so: C# and CLR and garbage collectors are not revolutionary or a technical leap over Java and JVM. The memory footprints will be huge like they are with Java once the benefit of running the CLR in the OS fades away. Does anyone know a C# application of caliber of say, Eclipse? How much memory does it take to run it - 100M, 200M? I am yet to see any evidence that C# would do any better than Java on the desktop when it comes to the end game.

    People assume too much. I am going to wait and see. Maybe I will get to call .NET a big elephant technology.
  108. Obviously you can not say what the state of Java is without comparing it to its competition.People are assuming that .NET will succeed on the desktop. Yes, it will succeed, but it will be a rewrite of old applications written in C/C++ with new C/C++ applications. Will you see C# applications? Is everythign going to be C#?My bet is that you will not see much of C# on the desktop at all, and this is why I believe so: C# and CLR and garbage collectors are not revolutionary or a technical leap over Java and JVM. The memory footprints will be huge like they are with Java once the benefit of running the CLR in the OS fades away. Does anyone know a C# application of caliber of say, Eclipse? How much memory does it take to run it - 100M, 200M? I am yet to see any evidence that C# would do any better than Java on the desktop when it comes to the end game.People assume too much. I am going to wait and see. Maybe I will get to call .NET a big elephant technology.
    You miss a point. When someone assumes to express some opinion about the evolution of the X or Y software technology he must not forget that IT as engineering consist from two inseparable components Software and Hardware – it’s about having vision over the field your profession evolves - and always the hardware will be the low level which will dictate how the software will evolve or behave or disappear :-) As long as in the maximum 3 year the standard desktop will have multicore chips and more than 1Gb available very fast RAM as standard installments, you must be a suicide-manager :-) if you don’t take seriously in consideration the development of desktop applications in managed environments. Another argument will be that Microsoft itself will enforce the usage of CLR for desktop apps because is the real unique way to solve their ancestral security problems, being Windows or POSIX systems( Linux/Unix’es), the common security problems appear because of memory access from user applications direct in the kernel or system shared/reserved memory, and Java must be able to offer the same level of usability and reliability if doesn’t want to get KO. See my previous post about what will mean this, it’s my personal opinion but I tried to be as accurate and logic as possible, of course we live in a free world and others can have different opinions, but don’t miss the big picture when you add your arguments.
  109. You guys are missing one important aspect of C#/.Net and it really displays a lack of a working understanding or knowledge about C#/CLR. What makes C#/.net so attractive is its seemless integration with regards to making system calls. Please let me deonstrate.

    using System;
    using System.Runtime.InteropServices;

    namespace DLLImportDemo
    {
       public class DemoBeep
       {
    //Call Win32 api.
        [DLLImport("kernel32")]
        public static extern bool Beep(int freq, int dur);


        public static void SystemBeep()
    {
    Random rand = new Random();
          for(int i = 0; i < 1000; i++){
              
               Beep(rand.Next(1, 100), 100);
     

    }
    private static void Main()
    {
    DemoBee.SystemBeeb();

    }
       }
    }

    Making system wide calls in such a complex adventure in java whereas .net has broken this down to one line of code:
        [DLLImport("kernel32")]
            public static extern bool Beep(int freq, int dur);

    I think the inherent powerful nature of this is obvious.

    -M
  110. [DLLImport("kernel32")] [...] I think the inherent powerful nature of this is obvious.-M

    Uhm - especially fo enterprises which invested a lot of many in non Microsoft solutions :). How can I deploy this on my Solaris, Linux or AIX boxes?

    To be serious - a lot of enterprises have choosen Java because it doesn't hook into one particular platform. Tell them to use .NET because it is easy to call Windows kernel32 Beep function :))).

    Artur
  111. Uhm - especially fo enterprises which invested a lot of many in non Microsoft solutions :). How can
    oops: s/many/money/g :)

    Artur
  112. This is precisely my ponint. A general ignorance about .Net. You can use the same DLLImport attribute in mono or any compliant CLR implementation.


    -M
  113. This is precisely my ponint. A general ignorance about .Net. You can use the same DLLImport attribute in mono or any compliant CLR implementation.

    DLLImport attribute means nothing if particular native Windows function is not implemented on other platforms. Most Windows specific features are not (and won't even be in the future) available outside Windows. Thats exactly my point.

    Talking about mono and .NET. For me mono is nothing more than next interesting useless environment. Everybody talks about mono and nobody actually usues it in production. I'm not aware about one real production system deployed on mono. Saing that mono could potentially replace Java is just funny (even more funny than PHP could replace java on server). No real deployments, no IDEs, almost no tools, half of the MS .NET is still unimplemented, not enterprise features nor WinForms, slow, in very early sage of development, almost no support from ISVs etc, etc. Anybody still wants to replace Java with mono?

    Artur
  114. "Saing that mono could potentially replace Java is just funny"

    So it is is it? In this very moment the Gnome guys are discussing replacing Linux "Java Desktop" with mono..When you use Mono and Glade they are using just the DLLImport attributes you bash..Why don't you try a "hello world" app in Mono Glade and compare it to a Swing application..

    hi hi

    Talk about not having a clue..

    Regards
    Rolf Tollerud
  115. "Saing that mono could potentially replace Java is just funny"So it is is it? In this very moment the Gnome guys are discussing replacing Linux "Java Desktop" with mono..

    Gnome guys discussing strange things all the time. Once they decided to include mc (midnight commander) into gnome desktop. They forced developers to stop developing (very good, light, extendible) command line version and switch to Gnome. Then they killed project. They changing their minds all the time. It's why I switched to KDE.
    When you use Mono and Glade they are using just the DLLImport attributes you bash..Why don't you try a "hello world" app in Mono Glade and compare it to a Swing application..hi hiTalk about not having a clue..RegardsRolf Tollerud

    Yes I tried - but what's your point? Are You saying that You can write hello world app in mono - it means that it could replace java?

    Artur
  116. "Saing that mono could potentially replace Java is just funny"So it is is it? In this very moment the Gnome guys are discussing replacing Linux "Java Desktop" with mono..When you use Mono and Glade they are using just the DLLImport attributes you bash..Why don't you try a "hello world" app in Mono Glade and compare it to a Swing application..hi hiTalk about not having a clue..RegardsRolf Tollerud

    I am wondering if sands are shifting. I see a whole lot more KDE stuff than I see Gnome. If Gnome is losing popularity, so maybe Mono is betting on the wrong horse? If the horse is shot from underneath the rider, the rider is left helpless. What is Novell going to do with it now when KDE proves to be a better setting with Suse Linux? Is the project doomed? Is the funding going to be over soon? Mono's days are numbered. FUD FUD FUD.
  117. QT# binding. You guys are out of touch.

    -McCorney
  118. Tero,

    "They have the luxury to wait for others to struggle to find practical solutions to their theories"

    But that is not how Sun works. Sun & JCP sit down by their desktop and constructs theoretical solutions and frameworks that is distributed from above, that is called Taylorism, not capitalism.

    Capitalism is about how a lot of practical solutions are fighting each others until the market decides, the opposite to Taylorism.

    Artur,

    "Yes I tried - but what's your point? Are You saying that You can write hello world app in mono - it means that it could replace java?"

    No. It is about writing Rich-Client applications with C#/Glade/Gtk that communicate with web-services, out-competing the Big Java Elephant Server solutions.

    Tero,

    "I am wondering if sands are shifting. I see a whole lot more KDE stuff than I see Gnome"

    Qt is bloated, in C++ (not C) and not 100% Open Source. Don't count on it. But a Qt# is in the oven too!
    http://qtcsharp.sourceforge.net/

    Regards
    Rolf Tollerud
  119. It is about writing Rich-Client applications with C#/Glade/Gtk that communicate with web-services, out-competing the Big Java Elephant Server solutions.Tero,"I am wondering if sands are shifting. I see a whole lot more KDE stuff than I see Gnome"Qt is bloated, in C++ (not C) and not 100% Open Source. Don't count on it. But a Qt# is in the oven too!http://qtcsharp.sourceforge.net/RegardsRolf Tollerud

    So let me get the story clear here. You're saying everything should be done with Rich clients. Including things like access to shared data that is used by multiple users, which need to propogate across hundreds and thousands of users with minimal (sub-minute) latency.

    There's no logical flaws in that reasoning at all. Everyone should just throw away database servers and write everything to flat files with rich clients. While we're at it, we should also throw away IIS, file servers and email servers.

    who needs mainframes, databases, and servers when you have rich clients. Let the new religion begin.

    </joke>
  120. Rolf Tollerud for president.[ Go to top ]

    Rolf is a revelation in the pit of theserviceside mire. All you guys are wally's and will be forced to use .NET/Mono because Java advocates have stuffed it up.

    3 cheers for Rolf.

    regards
    Peter
  121. Wow, is TSS under Microsoft attack?[ Go to top ]

    Rolf is a revelation in the pit of theserviceside mire. All you guys are wally's and will be forced to use .NET/Mono because Java advocates have stuffed it up.3 cheers for Rolf.regardsPeter

    the last time I checked, both client and server applications have uses. each one is good at certain types of applications. A good developer separates the BS from reality and figures out how to deliver high quality software to their customer. Making statements like "all you guys" might be a bit heavy handed.

    I would think, a good developer doesn't need to make silly statements like "big java elephant" or "webservices will kick java's butt." Mis-using technology has nothing to do with the value or flaws of a given piece of software. In my biased opinion, none of what Rolf says has any technical weight, and is roughly equivalent to marketing fluff. If a programmer falls for marketing hype, the individual is the one to blame, not the technology.

    If I get a trojan on my windows XP box because I didn't install a firewall and anti-virus, does that make windows a plague on humanity? Like wise, if java developers mis-use EJB doesn't mean EJB is evil or bad. If I try to reproduce Photoshop with .NET controls, does that make .NET a failure?
    Just like everything else, logic is a construct. In order to paint a vivid picture of success and failure, one has to first know how to use a brush and where to apply the pigments. Having a canvas, brush and paints doesn't automatically mean a person can reproduce "the last supper." It's easy to xerox copy a painting and claim to be a master.
  122. help is underway[ Go to top ]

    Peter: "Like wise, if java developers mis-use EJB doesn't mean EJB is evil or bad. If I try to reproduce Photoshop with .NET controls, does that make .NET a failure?"

    Peter, you are correct in this of course. Neither is there anything wrong with Java.

    But I despair of the fact that Java guys think that a person that blurbs "Organize inter-service transfers according to use cases from known domain objects into a coarse-grained Composite" is cool and is stupid enough to believe that he actually knows what he is talking about. Therefore it is a danger that all Java is "thrown away with the bath-water" because it could not be "washed clean from the Java culture".

    But of course Real Programmers like Rod Johnson and Juergen Hoeller are powerful cleaningmen with strong detergents. We'll see what happens.

    Regards
    Rolf Tollerud
  123. bad programmers are every where[ Go to top ]

    what I find rather useless is disparaging a community based on the feedback of a few individuals or marketing material. stupidly building an application without taking time to really understand the domain problem and the business needs is flat out incompetence. Have I done it before in the past? Definitely when I first started programming in the early 90's. It's not an easy lesson to learn, especially if someone believes that computer science theory is dogma to be followed at all cost.

    But to discredit an entire field of study, because organizations have a vested interest in promoting their brand of computer science isn't all that useful. Learning the soft/business skills is tough and it takes a lot of time. In my own experience, many people lack the people skills and often don't realize what the business guys say isn't what they mean (myself included).

    Decifering the real meaning of what an individual says isn't a skill most CS programs teach. I've never taken a college computer science course in my life, so I have no first hand experience. From the little I know of computer science courses, none of my friends were required take a class in creative writing, or debating. If we look at someone like Carmack, he is able to balance computer science theory with practical skills.

    I don't know about others, for me this is very hard to achieve and takes a lot of patience. Whether java, .NET, webservices or messaging survives as a fashionable platform is not material to me. Getting the job done with the highest degree of quality is all I care about. Sometimes that means taking time to do research and make sure I don't choose a bad algorithm. Sometimes it means post-poning theory/research until the entire team has a better idea of what the business requirements are. Discussing the specific details, merits, and flaws of a technology is a great learning process.

    Making overly generalized statements without specifics is useless. I'm guilty of it all the time. Even though I try to be reasonable and objective, it's just too easy to rant and over generalize.

    Back on topic, it's obvious Sun is getting beat by IBM in some areas and Sun does need to figure what their future is. I don't equate Sun with Java, nor do I equate J2EE with java. Therefore, I only use J2EE as a marketing term and don't consider it more than that. It's no different to what Microsoft did in the early days of .NET. The marketing guys felt is was a good idea to label everything ".NET", which ended up creating confusion. Once MS realized that, they changed their marketing plan. the more things change, the more they stay the same.
  124. All hail Troll Olferud[ Go to top ]

    And all us Wallys promise to snigger when .NET qualifies for a Darwin award.

    Honestly you get excited just because M$ produces a Java knock off that isn't a complete wreck on the first release. Even M$ are not mircle workers, give them time for a few more releases. (Not to mention that accepting an "open standard" from the Dark Master of the lock-in is a leap of faith that dwarfs 'trust in Sun')

    Those guy are masters, just look what happened with the registry - M$ took a brilliant idea to streamline configuration data and turned it into the font of all system instability.

    Back on topic as far as I can tell Java is fighting fit with a multitude of new projects in every imaginable application - I'm not a one language programmer, but even if I was I wouldn't be worried.
  125. How does rich clients solve server challenges[ Go to top ]

    So let me get the story clear here. You're saying everything should be done with Rich clients. Including things like access to shared data that is used by multiple users, which need to propogate across hundreds and thousands of users with minimal (sub-minute) latency.There's no logical flaws in that reasoning at all. Everyone should just throw away database servers and write everything to flat files with rich clients. While we're at it, we should also throw away IIS, file servers and email servers.who needs mainframes, databases, and servers when you have rich clients.

    Rich Client doesn't always mean Thick Client. I guess it depends where one draws the line - or defines each one.

    I've been doing Rich client (not Browser) for years and accomplished what you have described. Let the server continue to do what it does best. But serving up a GUI everytime someone logs in isn't one of them.
  126. good point[ Go to top ]

    Rich Client doesn't always mean Thick Client. I guess it depends where one draws the line - or defines each one.I've been doing Rich client (not Browser) for years and accomplished what you have described. Let the server continue to do what it does best. But serving up a GUI everytime someone logs in isn't one of them.

    I'd absolutely agree it makes sense to use rich clients for those cases. I don't think it would be necessarily less flexible either, if the rich client uses HTML or some other templating mechanism so the GUI can be updated without shipping a new binary.

    In that respect, a framework like XAML has potential of reducing the load on the front end webserver and improving the look and feel of the UI. Where I would hessitate using a rich client approach is where an application is targeted towards average users who do not want to install a rich client and prefer stock web interface.

    Applets tried to achieve that balance, but it never really took off. Instead, the web designers I know moved to flash. In terms of providing a rich client, flash does seem to provide the sweet spot for many of the web/graphic designers I know. Most of what they do is fancy animation/rollovers, so using something like .NET controls and applets annoys them to no end. Hopefully all this hype about webservices and SOA will die down, so programmers can get back to work.
  127. good point[ Go to top ]

    Having done VB, Powerbuilder, Html, Java, .Net and Cobol UIs - an html ui is less flexible and more difficult to maintain than a Java/.Net UI. Delivering a binary is the least of a developer's worry with JNLP and SmartClient. Sometimes a brower based UI is the lesser of two evils. XAML is somewhat better (Ok maybe much better). Scripting based UIs are such a pain.

    Hopefully all this hype about web apps will die down, so programmers can get back to work. :)
  128. seen plenty of abuse[ Go to top ]

    Having done VB, Powerbuilder, Html, Java, .Net and Cobol UIs - an html ui is less flexible and more difficult to maintain than a Java/.Net UI. Delivering a binary is the least of a developer's worry with JNLP and SmartClient. Sometimes a brower based UI is the lesser of two evils. XAML is somewhat better (Ok maybe much better). Scripting based UIs are such a pain.Hopefully all this hype about web apps will die down, so programmers can get back to work. :)

    on more than one occasion, I've been asked to write an application with graphic editing capabilities using javascript. of course I considered that abusing javascript and told the client it wasn't feasible. then again, I think most of this is just common sense. for simple stuff, HTML is quick and dirty. Even though news readers provide a richer interface, there's a lazy factor at work. Often, I don't want to open up another application because I already have a dozen applications and 20+ windows open.

    plus, I don't really care to have new posts d/l automatically. On the otherhand, if i had to write a blotter, it wouldn't make any sense to me to use HTML. given the complexity of a blotter, it much better to have a rich or heavy client. that's my own bias. I've seen some pretty impressive stock analysis applications written for applets, which were fast, beautiful and a pleasure to use. But to achieve that kind of quality takes super human swing/awt skills I don't possess.
  129. seen plenty of abuse[ Go to top ]

    "that kind of quality takes super human swing/awt skills I don't possess. "

    Oh come on Peter, the false modesty doesn't become you at all! :)

    Regards
    Rolf Tollerud
    (when I am modest, I will be perfect)
  130. seen plenty of abuse[ Go to top ]

    "that kind of quality takes super human swing/awt skills I don't possess. "Oh come on Peter, the false modesty doesn't become you at all! :)RegardsRolf Tollerud(when I am modest, I will be perfect)

    I kinda agree with Rolf on this one. If you can do all things you can do in the server-side arena, and I believe you when you say you cane, Swing/SWT/WinForms/etc. would/should be a breeze for you.
  131. I wish[ Go to top ]

    I kinda agree with Rolf on this one. If you can do all things you can do in the server-side arena, and I believe you when you say you cane, Swing/SWT/WinForms/etc. would/should be a breeze for you.

    I think you over estimate my abilities. If by breeze you mean working 80hours/week for 10 months, and asking a bunch of swing experts for help, than sure :) Over the last 4 years, I've gradually come realize just how little of swing and writing GUI's I know. When I look at tools like 3DSMax, Maya, Photoshop, Final Cut Pro and other polished applications, there's still a mountain of skills I would need to write something of that level.
  132. I wish[ Go to top ]

    When I look at tools like 3DSMax, Maya, Photoshop, Final Cut Pro and other polished applications, there's still a mountain of skills I would need to write something of that level.

    Ok. When you mention those tools, I understand where you are coming from. But 90 some percent of the time - not a problem.

    I got a couple of ex-mainframers/vbers going on Swing just recently. They really took after it. Didn't take that much work. And I wouldn't say I am a Swing expert. More like a Jack-of-all-Trades.

    If you need some help pointers, check out this blog. http://www.clientjava.com/blog/ Lots of good info.
  133. I'm just jealous of guys who wrote those apps[ Go to top ]

    Ok. When you mention those tools, I understand where you are coming from. But 90 some percent of the time - not a problem.I got a couple of ex-mainframers/vbers going on Swing just recently. They really took after it. Didn't take that much work. And I wouldn't say I am a Swing expert. More like a Jack-of-all-Trades. If you need some help pointers, check out this blog. http://www.clientjava.com/blog/ Lots of good info.

    even though I curse swing frequently, I do like it and it does what I need most of the time. I really wish I could go without sleep, that way I can write a game or an open source animation package. Not that there would be any point to it, since it would darn hard to match any of the apps i mentioned in the previous post.

    comparing swing to SWT, I prefer SWT. my definition of java platform is an open source stack including apache, objectweb, codehaus and eclipse :) I do consulting, so my perspective is completely bias.
  134. I'm just jealous of guys who wrote those apps[ Go to top ]

    my definition of java platform is an open source stack including apache, objectweb, codehaus and eclipse :) I do consulting, so my perspective is completely bias.

    Me too (although I am ok with Swing - depends on what I do). I do "consulting" too. But I am trying to write a product too. If only I could go without sleep... :) .
  135. maybe in 3 years I could[ Go to top ]

    "that kind of quality takes super human swing/awt skills I don't possess. "Oh come on Peter, the false modesty doesn't become you at all! :)RegardsRolf Tollerud(when I am modest, I will be perfect)

    honestly it wasn't false modesty. there's still a ton of swing stuff I don't know and have never touched. beyond making some simple custom graphs, custom table/tree models and simple gui utilities, I still consider myself a swing/awt light weight.

    given my current focus is server side, I really don't get much time to do GUI work. especially not anything complex or really difficult. I'm just too lazy to dig into it. Well more like I spend most of my free time reading server side stuff, research papers on distributed systems, rule engines, transaction processing and performance tuning.

    unless I switch focus to GUI's, for all intent and purposes I doubt I'd ever consider myself a swing/awt guru.
  136. MonodDevelop is way ahead of any development tool that came out with jdk1.0. Mono is way ahead of java/jvm that came out in jdk1.0 and it is rapidly gainning steam. Try creating a simple app with mono on linux and you see what I mean and create an equivalent in jdk1.5. It is quite embarrasing.
  137. You guys are missing one important aspect of C#/.Net and it really displays a lack of a working understanding or knowledge about C#/CLR. What makes C#/.net so attractive is its seemless integration with regards to making system calls.

    You mean the aspect that creates a vendor lock-in? Windows is a candyland of wonderful things for you to integrate in your application. If you do not want to leave this neverland, then that's just fine. I am just not one of those people.

    .NET's primary job is to create a vendor lock-in. Its job is to make it ever easier to get entangled to the Microsoft web. That is the only rule Microsoft never breaks: always tie everything to the Windows platform so that it supports the desktop monopoly. They never stray from this rule, ever.

    Why would I want to make a system call to Windows? My main incentive for using Java is that I do not have to. That is why I use Java, period.
  138. Yes why indeed?[ Go to top ]

    "Why would I want to make a system call to Windows?"

    But the hot thing right now is the make a system call to gtk! C#/Mono with Glade on Linux.
  139. My main incentive for using Java is that I do not have to. That is why I use Java, period.

    This is in part due to the fact that you guys really have a lack of working knowledge of c#/.net. You do not have to make system call if you do not want to. What .net has done is providing an intuitive way to do so if you wish.

    -M
  140. My main incentive for using Java is that I do not have to. That is why I use Java, period.
    This is in part due to the fact that you guys really have a lack of working knowledge of c#/.net. You do not have to make system call if you do not want to. What .net has done is providing an intuitive way to do so if you wish.-M

    I do not wish to do it.

    Why would I make an optional system call? That feature is not there to make optional calls but calls that can not be supported by the .NET platform (=legacy systems). If you are using that for any other purpose, I question your logic to do so.
  141. This is in part due to the fact that you guys really have a lack of working knowledge of c#/.net. You do not have to make system call if you do not want to. What .net has done is providing an intuitive way to do so if you wish.

    This is at a cost - cost of having OS dependent code. You and Rollf are claiming that mono could replace java. For me (and I thing a lot of people) important feature of java is portability - so easy way how we can potentially call native function is as /exciting/ as possibility of using GOTOs or having to deal with raw pointers. We - "java guys" - usually (there were some rare cases) haven't need to call native code, use GOTOs or deal with pointers. Why we should bother how easy we could possibly do these things we don't need/want?

    Artur
  142. You think you don't need them becuase Sun has convinced you that you don't and you have been living in a sand box for quite some time. That is the beauty of .Net. If you don't need system calls for a particular application then don't use it and stay within the CLR. If you find your self in a situation where you are saying to your self that "wow I wish I could make a system call" then it is available to you. I for one, a feature that I have grown to love is the System.Management namespace. Mono do intend to support that namespace sometime in the future.

    -M
  143. You think you don't need them becuase Sun has convinced you that you don't and you have been living in a sand box for quite some time.

    :))Yep and Dijkstra convinced me that using GOTOs is harmful - it's the only reason why I don't use GOTOs:)).

    I manage Unix development dept. in financial institution. We develop software for quite few years for different platforms (mainly various unice flavours). I force my people to write *portable* code - even in C. We just cannot write system dependent code - unless there is *very* serious reason for do so.

    Convince me why I should write native system calls?
    That is the beauty of .Net. If you don't need system calls for a particular application then don't use it and stay within the CLR. If you find your self in a situation where you are saying to your self that "wow I wish I could make a system call" then it is available to you.

    I found myself twice when I have/want to write system call. When you uses some feature twice per 5 years - belive me - it is not crucial how easy it is.

    Artur
  144. forget platform compatibility[ Go to top ]

    A windows app can be ported to Linux with the help of winelib for linux. Likewise Ximian provide a gtk installation kit that makes it possible to run Mono/Glade applications on windows.

    However, none of these solutions will be in any demand IMO.
    Remember when Microsoft tried to build Mac-Office with some libraries that allowed the same code to be recompiled for Macintosh? Do you remember the Mac users outcries? Do you remember how slow it was? In the end Microsoft had to rewrite everything special for Macintosh.

    A commercial desktop application must be written with native tools or it will be out-competed. And on Linux the C#/Glade combination is unbeatable, super fast apps with almost VB-like productivity.

    Regards
    Rolf Tollerud
  145. forget platform compatibility[ Go to top ]

    Rolf,

    You are way cool!

    -McCorney
  146. forget platform compatibility[ Go to top ]

    with almost VB-like productivity.RegardsRolf Tollerud

    VB apps are typically "productive" only in initial development. If they are complex or become complex, productivity drops.
  147. Sorry but, I don't know of what you are talking about.

    Clearly, the only reason that somepeople are using MS platform is because they don't want to choose of any solution. They want default solution.

    If you go and chalenge the solutions (this means you will compare for instance websphere soluion with MS .net solution) then this would be fair.

    But smart people that whant to introduce bias in recomandations, usualy chalange the Java platform agains the MS .net solution. Doing this is unfair has it does not address at all the same issues.

    Because you have to compare comparable items.

    Comparing the Java platform to the MS .net platform or comparing the Java platform to the mono platform (no kidding here plz).

    Doing that on both cases, Java platform wins on all the criteria (ROI, TCO, MiddleTerm vision, LongTerm vision, EOL perspective,...).

    An aside note, mono has nothing to do with MS .net platform, please remind this important statement once for all !
    Mono is intended to implement a subset of the .net features, and some of the .net behavior. So, in most of the cases running a .net project on mono, will just fail. And the other way is almost non-guaranteed either as there is no cross compatibility guaranty (MS keep the entire control of the .net spec, and can decide at any time to break any element that is not define at ECMA, which means almost 80% of the platform).

    As a consequence, choosing .net is a dead-end solution to Microsoft Windows OS. Because all the top manager from MS has made it clear, .net will only be ported to MS only OS (of course it is so, other, why will MS attack their own marketshares of their milking cow ?!?).

    Going to mono is also almost a dead end, because mono does neither offer a complete platform (such ones that both Java or .net offers) nor a complete solution (you will have to use studio .net and all other element from the MS solution set, all that are windows OS based ). So as a consequence, you will end up with either working with bare C/C++ code (having usual dev cycle and TCO ... and headaches), or will be constrainted to move to MS .net, and hence Windows again :(

    As a consequence, mono is not a middle term or long term solution that any project can trust as "risk free".

    As a conclusion, Java has nothing to fear from Mono.
    And about MS .net, the choice between .net and Java is no more on technical ground, but always political nowaday ;-)

    IMHO, having MS .net pursuiting Java is a very good point, because it push the core spec leaders to evolve, inovate and enhance the platform and not keep the "good old way" the default solution. This is why I think that on the long term trend it is MS that will make Java one the the corner stone technology of this century.

    Personal mind : The only thing that are missing right now, is that FSF commit to the Java platform take seats at the JCP put ressources to the GNU Classpath project and provide us a certified J2SE 5.0 GPL implementation. This will once for all remove the dependancy on OS for tomorrow softwares.

    Personal hope : I am very excited about the Looking Glasses Project (LG3D) and I realy think this is the next big thing that could bring a complete new momentum to java on the client side if used with the webstart (JNLP).
  148. Lots has been said in this thread, so I'll simply make my points and go on my merry way.

    - I work for a firm that does enterprise HR software for the US Navy. Our current application supports 600,000 users, concurrent has been seen up to 10,000. We run on 4 dual-xeon boxes with a totaly of 12 JVM instances clustered on oracle 9iAS. We access 3 different disconnected databases (not ours), one of our own, and two web services providers. We maintain a 3 month release cycle with mind-boggling accuracy.
    - Based on our success with this application, the US Airforce has contracted with us starting at $29 mil for 4 years, it will be 100% java. (ok maybe .003% javascript ;) )
    - now, THAT is what I call enterprise software. I challenge anyone to try that with PHP.
    - I have worked for a company that attempted enterprise document management in PHP, it was a total mess. It was also 50k lines of sourse and had less functionality that the above mentioned application that is just over 30k lines.

    If Java is good enough for the most advanced military in the world, its good enough for me ;)
  149. Java dying??/ I think not.... follow the $$$[ Go to top ]

    Kurt, thanks for the excellent info.

    On the other hand, having spent 8 years in the AF working on the most modern systems (aircraft and satellite systems) I've seen what is good enough for them. :) And what is important to them. "I know you are presenting to me a Powerpoint presentation on the status of a > billion dollar satellite but you need to use colors I like on it first."
  150. Java dying??/ I think not.... follow the $$$[ Go to top ]

    Haha Mark, I totally agree. We also have an excellent demo/presentation team ;) Once the foot is in the door though, they do have very strict technical requirements and I don't feel we could have the success that we do without Java and J2EE.
  151. no problem with HR application in PHP[ Go to top ]

    Yes, Pentagon is a good customer. We know that they always 5-6 times as any other company for everything!
  152. pay, that is! :)
  153. Yes, Pentagon is a good customer. We know that they always 5-6 times as any other company for everything!

    you wouldn't believe how much the government overpays and how much paper gets pushed around. try ordering a microscope and most likely it will take 12months. the best part is, you'll have to fill out 20 different forms in quadruplicate. should a form be incorrect or have a typo, god help you.

    sorry for the off topic response, but I couldn't resist.

    on the java platform question, just because the government pays too much for everything, that doesn't necessarily validate or invalidate technology X. I've heard horror stories from CSC people about lotus notes, exchange and just about every product the military uses. I would think the military choosing java or J2EE doesn't prove anything, other than IBM's marketing appears to work on the government.
  154. if you're lucky it's just 5-6x more $$ wise[ Go to top ]

    Yes, Pentagon is a good customer. We know that they always 5-6 times as any other company for everything!
    you wouldn't believe how much the government overpays and how much paper gets pushed around. try ordering a microscope and most likely it will take 12months. the best part is, you'll have to fill out 20 different forms in quadruplicate. should a form be incorrect or have a typo, god help you.sorry for the off topic response, but I couldn't resist.on the java platform question, just because the government pays too much for everything, that doesn't necessarily validate or invalidate technology X. I've heard horror stories from CSC people about lotus notes, exchange and just about every product the military uses. I would think the military choosing java or J2EE doesn't prove anything, other than IBM's marketing appears to work on the government.

    I know we all think that the Gov pays too much. But I know people who won't do business with the government because a.) it doesn't pay well and b.) they don't pay on time.

    Sorry for my silly comments but I think Kurt's point was that they (not the gov) had a VERY successful complicated project with Java and now the military wants them to do more. Usually they don't care and always go to the lowest bidder. Experience or no experience.
  155. the users against the mainframe guys[ Go to top ]

    Wow, this certainly is a fun fight! And the best thing with it is that it is almost a perfect copy of what happened before.

    I started in the computing business with Macintosh, it was before windows and we used tools that was named "Omnis" or "4th Dimension" or something like that. The users loved us and the middle management CTOs hated us. One memorable week 7(seven) middle management guys at Philips resigned in anger!

    "Those wore the days". (nostalgic..)

    But I never though it would actually happen again! :)

    Regards
    Rolf Tollerud
  156. if you're lucky it's just 5-6x more $$ wise[ Go to top ]

    Thanks Mark, That was my point exactly. We've been outbid before, but they keep coming back to us due to the quality of our software. They are even recommending (forcing) some of their other vendors to use Java due to the lower maintenence and faster turnaround times.

    Agreed, they have no idea how to spend money (eg. I just spent a week in a oceanfront marriott in virginia beach on the Navy's dime for meetings I didn't really have to be at ;) )
    BUT they are getting stricter on large spending, heck the Navy budget was so screwed up they reecently had to cancel all cable television to the barracks. Along with all kinds of other acceptable comfort expenses.

    In IT they're looking for lower TCO all across the board, even if that means slightly more expensive upfront costs for systems. Remember, 80% the cost of software is often the maintanance. Java helps greatly here because of the robust frameworks available and strong typing. Generics while somewhat syntactic sugar, will greatly enhance maintanability as well.

    Cross platform, scalability, low maintanence, flexibility, and massive amounts of 3rd party tools and APIs. You just cant find all of that in PHP or .NET.

    I get the feeling this was one of those TSS 'lets incite an argument to make the site more active' stories ;)
  157. good way to put it[ Go to top ]

    I think you hit the nail on the head there :) nothing like a inflamatory thread to boost activity.

    it's funny how things vary even in the military. just goes to show, even really large entities have multiple heads going in different directions. though I would credit the success to your diligence and not necessarily java. I'm sure given some other language, you could do just as good of a job.
  158. good way to put it[ Go to top ]

    just goes to show, even really large entities have multiple heads going in different directions.

    You have no idea my friend ;) Lemme tell you, military politics and technology do not mix. There are so many systems in the navy that overlap and all controlled by different commands and nobody wants to give up thier data/power (data==power). Luckily there are a few 2 and 3 star admirals with vested interest in the project we are involved in (our app is a small part of huge IS and training overhaul in the navy) and they have a solid unified vision and the power to make things happen. Of course that does not get us out of day to day bullcrap.

    Getting off topic here, but to tie it back in... Java greatly helps our ability to remain agile in the quickly changing environment that is Naval IT. Dilligence helps for sure, and we have excellent project management staff, but there is no doubt that Java is the core facilitator to project sucess. 3rd party APIs included of course, I'd like to thank Struts, Hibernate, ibatis, Eclipse IDE, and Ant; without them and projects like them, enterprise Java would be another beast entirely I think.
  159. please wake me up[ Go to top ]

    There are different grades of being bored. We have ordinary boring, deadly boring, piercing boring, etc. But to cite projects examples as a argument to which develop environment is best! Do you want me to begin quoting from my bank of 300-400 large .NET projects? Please spare me.

    For that we have to invent some completely new and original adjective to boring.
  160. if you're lucky it's just 5-6x more $$ wise[ Go to top ]

    BUT they are getting stricter on large spending, heck the Navy budget was so screwed up they reecently had to cancel all cable television to the barracks.

    Bet they still have cable in the Officers' Club. Grrr.

    Guess the sailors will have to entertain themselves. Uh-Oh. :)

     
    In IT they're looking for lower TCO all across the board, even if that means slightly more expensive upfront costs for systems. Remember, 80% the cost of software is often the maintanance. Java helps greatly here because of the robust frameworks available and strong typing. Generics while somewhat syntactic sugar, will greatly enhance maintanability as well.Cross platform, scalability, low maintanence, flexibility, and massive amounts of 3rd party tools and APIs. You just cant find all of that in PHP or .NET.

    Preach it brother! I am facing this right now in a couple of projects. I could save time AND money if I was able to use Java and able to use all that surrounds it.