Home

News: Design-Time API Promises to make Java more like VB

  1. Artima has published a short article describing the Design-Time API for JavaBeans, which was recently approved as JSR 273. This API promises to bring VB-like ease to Java development.

    In the article, Joe Nuxoll, spec lead for JSR 273 and two-time winner of the Java superbowl, says that "the lack of a vibrant Java component market may partly be the result of a cultural gap....In the Java world, the solutions are focused on class libraries, not component libraries. .NET developers, on the other hand, think of solving problems in terms of components, rather than class libraries and configuration files."

    Read the article here:

    http://www.artima.com/articles/designtime.html

    Do you believe that a vibrant JavaBeans component market can ever take off in the Java community? Also, do you feel VB developers could be persuaded by better, more VB-like tools, to switch to Java?

    Threaded Messages (226)

  2. Just like Delphi[ Go to top ]

    One of the things I remember from Delphi (the tool serious developers used instead of the VB toy-thingie) is that you could have some modifiers so that you might compile a runtime-only version of your component/library.

    We could even see people freely distributing components with limited design-time behaviour (you would have to pay to get the full version).

    Design-time behaviour is VERY desirable for UI components, for example. I'll check it out...
  3. Do you believe that a vibrant JavaBeans component market can ever take off in the Java community? Also, do you feel VB developers could be persuaded by better, more VB-like tools, to switch to Java?

    No. There are plenty of extendable Java Components in the Open Source world. And there are enough closed source ones too.

    I switched from VB to Java (VAJ at the time). The Java IDEs are pretty close to what VB had. And in some ways it is better. Custom properties is the only thing missing, that I can think of - and how many people really used them.
  4. I think they misunderstand the problem. Why Java Beans components never took off.
    Borland has quite rich set of Java Beans components packaged with JBuilder.
    Other companies could also create such component libraries.
    I mean technically there was no problem.
    But markets do not born and strive because of existance of particular technlogy.
    There has to be demand for components in java world...and it is not there.
    Why ?
    Because main destination of components is building visual GUI. And since java sucks as a client side programming, majority of developers prefer develop client side applications on other, more suited for the task technologies, VB, dotnet, delphi, or web GUI like ASP, JSP, PHP etc.

    So as you see - there's no demand.
    And they can approve as many JSRs as they like. That wont change the demand, until they fix java on client side.
    In other words make java clients installation as smooth and painless as possible.
    Now where's JSR for that ? :))
  5. And since java sucks as a client side programming,
    Hmm. It doesn't for me and many others.

    Why it didn't "take off":
    1. Timing. Everyone (well almost everyone) wanted to do Web dev when Java was coming into its prime.
    2. Have you ever tried to create a component in VB? Yuck. Ever tried it in Swing? Lots easier.
    3. You get more with Java than you did with VB. (One big one - layout managers)

    VB is (or was) not great for creating client-side programming. It was great for quick and dirty.
  6. competing with VB 3.0[ Go to top ]

    And since java sucks as a client side programming,
    Hmm. It doesn't for me and many others.Why it didn't "take off":1. Timing. Everyone (well almost everyone) wanted to do Web dev when Java was coming into its prime.2. Have you ever tried to create a component in VB? Yuck. Ever tried it in Swing? Lots easier.3. You get more with Java than you did with VB. (One big one - layout managers)VB is (or was) not great for creating client-side programming. It was great for quick and dirty.

    Hi,

     I've long thought about it and discussed with fellow developers on what would make
    a java IDE to be atleast as productive as VB 3.0 was. Today, I'm ashamed to say that VB 3.0 could be learnt in half a day AND a newcomer could develop a small 10 form/window application in a week's time.

      Of course, MS extended VB's ease of use to Visual Interdev.
     
      The day a JAVA IDE lets you develop a 10 form/window app in 7 days, that is the day that Java's popularity will really rocket. It is important that the application developer should be able to learn it in NOT more than 3 days.
     
    Unless the model that you're working on will also help doing something similar to developing app for the 'net (like Visual Interdev) where
    the developer who's used to the VB/Rave model of developing can reuse his development experience, there won't be big adoption/popularity.
     
        I hope that all the popular Eclipse/NetBeans/Swing/SWT/IDEA authors latch onto this and there is effort to have them interested in this technology. So, when the JSR is released/published, there will be enough books about the technology and upcoming IDEs supporting this technology.

    Those interested should start participating in this JSR as individual community members -- found that you can participate in the JCP as an individual for free.

    P.S : I have been developing software for a dozen years now. started using Borland C++, OWL
    and then later VC++/VB, etc and the last 6 years been doing Java -- mostly server side.

    BR,
    ~A
  7. I have witnessed a lot of failed attempts at developing complex applications with VB. The problem is that managers see a 10 form application produced in record time and then decide that they need that kind of productiviy. For release one it feels like they are geting that kind of productivity, but then eventually they hit the complexity limit of the product and the project slowly begins its decline. I have never seen a well architected, easily maintained VB app. Every one I have ever seen is a bunch of hacked together procedural scripts inside a UI. This works great for very simple apps, and sucks for complex apps. So I guess productivity depends on the kind of project you are working on
  8. simple apps fast not good[ Go to top ]

    Speed can be a critical factor in development. VB code can be written to be maintainable, or not. Believe me, I've seen plenty of unmaintainable Java code. The thing is, if you can go fast, you can get a foothold in a market, where your well architected brethren starve to death trying reach perfection.
  9. simple apps fast not good[ Go to top ]

    VB code can be written to be maintainable, or not.
    The VB IDE makes it very difficult (right next to impossible) to have maintainable code.
    Try renaming a class or form.
    Try having a large project and then page through the object list (no way to group other than by type).
    Try having a form be both and MDI child and a modal form.

    If VB(classic) was so good, Microsoft would not have dropped it for VB.Net.
  10. VB.net != bad VB[ Go to top ]

    VB code can be written to be maintainable, or not.
    The VB IDE makes it very difficult (right next to impossible) to have maintainable code.Try renaming a class or form. Try having a large project and then page through the object list (no way to group other than by type).Try having a form be both and MDI child and a modal form.If VB(classic) was so good, Microsoft would not have dropped it for VB.Net.

    Even though I don't like VB. I think the argument that VB.NET exist and therefore VB may be bad isn't necessarily a valid or good argument. MS has to come out with new versions of VB to make money. Regardless of how good or bad VB was, VB.NET does not provide credible proof to me. Since there are opinions for and against VB.NET, it would be hard to prove VB.NET is a clear winner. But really, I don't think concensus is a good way to measure either. I'd lean towards a case-by-case evaluation of whether or not VB was sufficient.

    peter
  11. VB.net != bad VB[ Go to top ]

    Regardless of how good or bad VB was, VB.NET does not provide credible proof to me.
    Maybe not. But it should make you go HMMMM.

    Maybe I should have said VB.Net/C#.

    I am not saying that VB being "bad" is the only reason that they created .Net. They could have just enhanced VB it was really that good. The problem with VB is that it can be separated from its IDE. They go together.
  12. VB.net != bad VB[ Go to top ]

    CANNOT be separated. Oops.
  13. simple apps fast not good[ Go to top ]

    Some people are not happy MS dropped classic VB:
    http://classicvb.org/

    Joel on Software's quote at the bottom of the page sums it up well enough.

    Regards,
    Henrique Steckelberg
  14. simple apps fast not good[ Go to top ]

    Some people are not happy MS dropped classic VB:http://classicvb.org/Joel on Software's quote at the bottom of the page sums it up well enough.Regards,Henrique Steckelberg

    Yeah I know. I was on some VB boards a few years ago. Major wailing. Some people just don't want to think or learn new ways. People will be happy to know that VB.Net is stinky in its own right. Some things I hate about the VB IDE have been expertly incorped into VB.Net.
  15. competing with VB 3.0[ Go to top ]

    The day a JAVA IDE lets you develop a 10 form/window app in 7 days, that is the day that Java's popularity will really rocket. It is important that the application developer should be able to learn it in NOT more than 3 days.

    That sounds exactly like Java Studio Creator to me.
  16. Java studio creator[ Go to top ]

    Somebody compared ease of use with the Java Studio Creator.
    I had a look at it, it has potential, but in its current state it is close at the edge of uselessness, part of it is to blame for a lack of standardized component description we discuss here.

    Studio Creator has two big problems:
    a) It reduces itself to the most basic JSF components, no extended ones. Which means, it is good for simple hello world programs, but you run into JSF user control coding hell (and believe me JSF control coding is hell, with too many code artefacts for one control, just like every jsf coding without tool support is hell) if you need more than the few standard basic html controls.
    You cannot really add third party control sets to the tool, because you cannot display them visually, because up until recently there was no real spec on how the Creator handles the controls in the UI builder, and there still is no standard.

    The second biggest problem is its total reliance on JDBC and automated JTA for transactions, while most webapps currently are done with the defacto standard hibernate. Although we will probably never see Hibernate support in it, supporting JDO would even help to get out of the JDBC hell this product is in.
    I will give a straigt example:
    You try to code a simple master view: You layout the form
    a table control is there:
    Within 5 minutes you run into the classical problem paging:
    No paging support from within the control, you start to search on the net and you can find various libs which add this control, wham you run into the next problem, no visual layout of the controls, although you can reach them within the editor.
    Next problem, you dock the table and the query to the control, next problem the gluing code docks a result set onto the control, that basically means no control over your transactions and open connections. I am not sure (since I dont know jta and the code of the reference implementation too much), but I have the eery feeling the connections are kept open until timeout that way, which is a no go once you have more than a handful of users.

    So basically you run into very similar problems with the creator as in Visual basic (dread awful api for user interface element creations, you run into limits if you cannot revert to third party controls, the application seems to have scaling problems on the db side, you run easily into db dependency hell thanks to the usage of plain jdbc)
    But that is mostly the fault of the Creator itself, which has nice half baked concepts but lacks severely in the execution departement.
  17. Java studio creator[ Go to top ]

    Somebody compared ease of use with the Java Studio Creator.I had a look at it, it has potential, but in its current state it is close at the edge of uselessness, part of it is to blame for a lack of standardized component description we discuss here.....

    You make a lot of good points. I would not go as far as to say it is that useless, as I have seen people happily develop apps with it, but I do agree with a lot of what you say, particularly about the lack of Hibernate or JDO integration.
  18. Java studio creator[ Go to top ]

    I'm a little confused -- the Data Table does offer paging. Just bring up the Table Layout customizer on the table and switch to the paging tab.

    Also, many third party components -can- be imported directly into the tool and get visual WYSIWYG designtime editing. For example I know Otrix, SoftwareFX and ESRI have Creator add on components that give you full designtime fidelity - these were demoed at JavaOne last year.

    Regarding JDBC -- you don't have to use JDBC directly; you can add your own middle layer. I blogged on how to do that here:
    http://blogs.sun.com/roller/page/tor/20050423#updatable_data_table

    It's true that "there is still no standard" -- that's PRECISELY the purpose of the JSR that we're discussing in this thread!!
  19. Thanks for the reply[ Go to top ]

    I must have missed the in built paging, sorry for the false statement, but that does not really make my other statements invalid, I am glad that things are moving forward with a standard design time api (hence the Creator discussion is not totally off topic), that is badly needed for creator.
    True some third party component packs can be plugged in, because the Creator team recently has announced the design time API, but some are not there yet.
    As for the missing JDO/Hibernate integration, I know that you can bind those things over the JSF API, been there done that, but believe me the task is not trivial at all, you have to implement your own datamodel, which sort of is layed out for keeping the connection open and fetching one dataset at a time
    (I ended up with the implementation of a connection buffer which keeps one page at a time and refreshes the page once the page boundaries are hit, which could be more easily solvable by the Hibernate cache itself, if I had more control over the transaction demarcation)

    But once you run into this and do it yourself you lose half of the really great functionality of the creator, the automatic binding of the forms and controls to the tables and the queries, which I really love.
    Dont get this wrong, the Creator has really substance, but it is not fleshed out yet fully. Adding more standard controls like a date control, html edit (all the niftyness, component packs like myfaces have) would help a lot, also the direct support of JDO and hibernate with a clearer transaction demarkation, so that you really can rely on being able to handle more than a handful of users.

    I am sure things will move there soon, but as usual, you probably have to wait for Version 2.0 or 3.0 like with every other product.
    (I have yet to see one JSF product within the 100-500$ pricerange which was totally satisfying to the needs of someone who develops enterprise apps, all of them had major problems which are solvable)
  20. competing with VB 3.0[ Go to top ]

    speed != productivity
  21. And since java sucks as a client side programming,
    Hmm. It doesn't for me and many others.

    I totally agree. I find client-side GUI development to be very simple and fast with NetBeans.
    VB is (or was) not great for creating client-side programming. It was great for quick and dirty.

    Yes, and I have seen some of the mess that can result when large projects are developed with it.

    Regarding the original post, the reason that VB (at least pre-.NET) users developed with components rather than class libraries was because you could not even write class libraries in VB - the language had no inheritance.
  22. Have you ever tried to create a component in VB? Yuck. Ever tried it in Swing? Lots easier.
    Are you serious?

    Creating a component in VB is a breeze. Note that I am talking about the *packaging*, not the API, which is what really defines a component.

    Write your code in VB, export a DLL or an OCX and the rest of the planet can use it. No need for them to recompile it, build it, update it or whatever. Just take a look at the IDL interface and you can invoke it from any language.

    HTML renderer, ftp object, graphic controls, etc... there are hundreds of these out there and I bet quite a few on your own machine, which you can discover and reuse with the OLEBrowser.

    Reusable Swing components?

    None.

    --
    Cedric
  23. Write your code in VB, export a DLL or an OCX and the rest of the planet can use it. No need for them to recompile it, build it, update it or whatever. Just take a look at the IDL interface and you can invoke it from any language.HTML renderer, ftp object, graphic controls, etc... there are hundreds of these out there and I bet quite a few on your own machine, which you can discover and reuse with the OLEBrowser.

    And managing versions of these on a Win32 machine can be a nightmare.
    Reusable Swing components?None.-- Cedric

    Wrong. There are plenty. A quick google shows this.
  24. Reusable Swing components?None.-- Cedric
    Wrong. There are plenty. A quick google shows this.I have quite a few Swing applications installed on my machine. How many components from these applications can I reuse?

    None.

    And by reuse, again, I mean I shouldn't need the source, nor recompile it. I should be able to look up all the various subcomponents "exported" by these applications and invoke them directly from my Java application.

    How many of these components do *you* have on your machine?

    I bet none too, simply because Java has no standard for this.

    On the other hand, I took a quick look at the controls I have on my machine (not even all written in VB, language is irrelevant for this) and there several hundreds that I can incorporate in my application in just a few minutes.

    The best part of it is: I didn't even install any of these components, they are just the building blocks of applications I am using.

    --
    Cedric
  25. Again, with proper quoting...
        Reusable Swing components?None.-- Cedric

    Wrong. There are plenty. A quick google shows this.

    I have quite a few Swing applications installed on my machine. How many components from these applications can I reuse?

    None.

    And by reuse, again, I mean I shouldn't need the source, nor recompile it. I should be able to look up all the various subcomponents "exported" by these applications and invoke them directly from my Java application.

    How many of these components do *you* have on your machine?

    I bet none too, simply because Java has no standard for this.

    On the other hand, I took a quick look at the controls I have on my machine (not even all written in VB, language is irrelevant for this) and there several hundreds that I can incorporate in my application in just a few minutes.

    The best part of it is: I didn't even install any of these components, they are just the building blocks of applications I am using.

    --
    Cedric
  26. Again, with proper quoting...
        Reusable Swing components?None.-- Cedric
    Wrong. There are plenty. A quick google shows this.
    I have quite a few Swing applications installed on my machine. How many components from these applications can I reuse?None.And by reuse, again, I mean I shouldn't need the source, nor recompile it. I should be able to look up all the various subcomponents "exported" by these applications and invoke them directly from my Java application.How many of these components do *you* have on your machine? I bet none too, simply because Java has no standard for this.On the other hand, I took a quick look at the controls I have on my machine (not even all written in VB, language is irrelevant for this) and there several hundreds that I can incorporate in my application in just a few minutes.The best part of it is: I didn't even install any of these components, they are just the building blocks of applications I am using.-- Cedric

    I have no doubt you know far more about this than I do, but I do re-use Swing components without re-compilation. I just make a JComponent subclass, put it in a Jar, and add it to my NetBeans component palette. That is it.
  27. Again, with proper quoting...
        Reusable Swing components?None.-- Cedric
    Wrong. There are plenty. A quick google shows this.
    I have quite a few Swing applications installed on my machine. How many components from these applications can I reuse?None.And by reuse, again, I mean I shouldn't need the source, nor recompile it. I should be able to look up all the various subcomponents "exported" by these applications and invoke them directly from my Java application.How many of these components do *you* have on your machine?

    Come on Cedric, then you most be doing something wrong...

    Let me count my latest swing application: I use JCalendar, ImageButton, JTableForEdit, Flex1Button, AnimatedCardLayout, NeonBorder, ShadowBorder, JPanelWithBackgound, JCheckBox3 (three state checkbox) and WatermarkViewport.

    I say those constitute components.
  28. I was involved in VB3 in the beginning. We used it as a way to get away from the 4GL environments were the database and programming language was hopelessly intertwined.

    Ah, VB3 with C ODBC API! What speed! What freedom! And the productivity compared to C++! Here you really could talk about driving rings around your opponent.

    If you knew your Charles Petzold you could work miracles with it.

    Reards
    Rolf Tollerud
  29. P.S.[ Go to top ]

    You should have seen the face of our presumptive customers when we opened an other database with "file/Open"! :)
  30. Ah, VB3 with C ODBC API! What speed! What freedom! And the productivity compared to C++! Here you really could talk about driving rings around your opponent. If you knew your Charles Petzold you could work miracles with it.

    Charles Petzold? If you really knew Windows programming, you'd know that the Petzold book was complete rubbish.

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  31. ?[ Go to top ]

    Ha ha, we are disrespectful today aren't we?

    We are talking about *the Charles Petzold*, the famous author of "Programming Windows" of five editions are we not? I though "Look it up in Petzold" remains the decisive last word in answering questions about Windows development. At last is was for 10 years ago when I was using it..

    We are obviously back in the old space again then, about the attitude of Java developers. Throw shit at Charles Petzold! Comparing some morons at Sun with one of the worlds greates geniuses!

    What’s happen to you? When did you totally lost yor sense of proportions? Weblogic 30X faster that Tomcat! :)

    I wonder whom you will take on next. The Pope? Jesus Christ? God?

    Call my up and I will recommend some good physiatrist.

    Madre Mia
  32. ?[ Go to top ]

    Throw shit at Charles Petzold! Comparing some morons at Sun with one of the worlds greates geniuses!

    Rolf, if you think that Petzold is a genius, then you've pretty much explained the extent of your technical capabilities, as if you hadn't already made your lack of ability abundantly clear.

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  33. Actually I was comparing the poor souls at Sun with Walter Bright, a bona fide genius or what do you think? One of your famous misunderstandings again?

    "How to keep it from becoming a mess is a different question"

    If I one day decides to compile a collection of "Cameron’s famous remarks at TSS", the you will see how I can make good use of the English word mess, I am sure that you are quite alone in having made so many stupid and incorrect remarks. Not to speak of your un-English behavior.

    Regarding Charles Petzold, why don't we comparing him to the staff at Sun? You know that for a long time they hired people without a proper screening etc. But by sheer coincidence (mistake?) they probably hired some good people. However they all have left now! What you want to call the people left working at Sun? Care to venter a suggestion?

    But what is important is not to compare single persons, but company with company! Try to compare Sun with Microsoft to get your proportions right. :)
  34. Rolf "AutoCAD" Tollerud[ Go to top ]

    I am sure that you are quite alone in having made so many stupid and incorrect remarks.

    This from the guy who said 'AutoCAD is OOP's ancestral beginning' and 'EBay is the most expensive IT project in history'. You never let facts or reality get in the way of giving us a good laugh.
  35. Reality is too boring[ Go to top ]

    I am sure that you are quite alone in having made so many stupid and incorrect remarks.

    This from the guy who said 'AutoCAD is OOP's ancestral beginning' and 'EBay is the most expensive IT project in history'. You never let facts or reality get in the way of giving us a good laugh.

    I think it's quite entertaining to read Rolf's comments. It's fun to see someone claim, the same query against a database with 50K rows will be just fine for a database with 10million rows. It's always entertaining to ask a technical question and get a non-sense quote from some obscure roman philosopher. Life would be way too boring without some spice.

    peter
  36. Actually I was comparing the poor souls at Sun with Walter Bright, a bona fide genius or what do you think?

    I have not commented on Walter Bright. I am sure he is bright, no pun intended, but I've never heard of him before this thread.
    If I one day decides to compile a collection of "Cameron's famous remarks at TSS", the you will see how I can make good use of the English word mess, I am sure that you are quite alone in having made so many stupid and incorrect remarks.

    Other than insulting ninjas by calling you a ninja, could you please point to an incorrect remark that I have made, and also the proof that it is incorrect?
    Regarding Charles Petzold, why don't we comparing him to the staff at Sun?

    Charles Petzold is an author of an introductory level programming book. What does he have to do with anything?
    You know that for a long time [Sun] hired people without a proper screening etc. But by sheer coincidence (mistake?) they probably hired some good people. However they all have left now! What you want to call the people left working at Sun? Care to venter a suggestion?But what is important is not to compare single persons, but company with company! Try to compare Sun with Microsoft to get your proportions right. :)

    You are a pathetic fanboy. I don't care to waste my time arguing about two big companies that I don't work for or invest in. If you want to play fanboy, go to the Anandtech forums and argue about AMD versus Intel, or Nvidia versus ATI.

    Look, you've already proven your complete technical ignorance. Don't insult yourself in public any more than you already have.

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  37. <sarcasm>
    Who is this Rolf that people speak of? Does he post on this site? I think I've read every posting that have made sense. I don't recall ever seeing his signature associated with any of them.
    </sarcasm>

    John Murray
    Sobetech
  38. "Look, you've already proven your complete technical ignorance. Don't insult yourself in public any more than you already have"'

    And you have yet again presented yourself as a dishonest person. I know that I any time can write an application that beat the hell out of your. That is valid for all aspects, speed, user functionality and pure beauty. That I can prove anytime. Please take up the challenge. I produce faster. I produce smaller.

    You are different from the others because you know that Weblogic is not 30x faster that Tomcat and all the other shit you have said about EJB and elephant containers and so on. You are a true genuine dishonest hypocrite.

    I know that you are only protecting you investment and want to go on selling "the worlds most expensive hash map". But to degrade yourself to that extent is really too much. Don’t you have any dignity man?

    "You are a pathetic fanboy"

    On the contrary, unlike you I persevere the truth
  39. "Look, you've already proven your complete technical ignorance. Don't insult yourself in public any more than you already have"'

    And you have yet again presented yourself as a dishonest person. I know that I any time can write an application that beat the hell out of your. That is valid for all aspects, speed, user functionality and pure beauty. That I can prove anytime. Please take up the challenge. I produce faster. I produce smaller.

    If you need the URL, it's http://forums.anandtech.com/ ..
    You are different from the others because you know that Weblogic is not 30x faster that Tomcat and all the other shit you have said about EJB and elephant containers and so on.

    Temper, temper ..

    Please quote the section in which I said "that Weblogic is 30x faster that Tomcat" and then disprove the information that I provided. It is on the public record, is it not? What could be easier for you to do than to quote what I said and prove it wrong?
    "You are a pathetic fanboy"

    On the contrary, unlike you I persevere the truth

    http://forums.anandtech.com/ .. go get 'em tiger!

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  40. But where's the facts?[ Go to top ]

    On the contrary, unlike you I persevere the truth

    Cameron has on numerous occasions given solid technical answers. He has demonstrated experience and depth in several areas. If you believe any statement made by any member of the community to be false or inaccurate, then simply show concrete proof.

    I've yet to see you provide concrete proof you know how to write high availability and high performance applications. You may consider writing rich clients for a small 50-100K row database to be the same as writing an application to handle 100K bulk transactions in one day, but they are clearly different. The techniques employed to solve these two different classes of problems are very different. Anyone that has actually work on these two classes of problems know this first hand.

    Honestly, go read the articles on MSDN that address HAV applications. They clearly state how different these problems are. Or are you saying that even MSDN is wrong and that Bill gates and Microsoft are lying about the difference between a rich client and a HAV server application?

    peter
  41. In the three years Cameron and I have discussed –

    1) Cameron bitterly defended EJB, now Spring and other light frameworks rules.
    2) Cameron said "nobody uses it" referring to .NET, now .NET has surpassed J2EE in new projects.
    3) Cameron said that stateless server applications "are no real applications", statesness is now in fashion of the day with SOA.
    4) Cameron say that "scripting" languages like Perl and PHP can not handle big sites, now livejournal.com with 50 million hits a day has a better uptime than TSS.
    5) Cameron lied recently* (July 16, 2004) that Tomcat used to be 30x slower than Weblogic
    6) Cameron believed Oracle when Oracle claimed they had turned "28x slower" to "28x faster" with a few tweaks of the code.
    7) Cameron claims that Java is faster that C# :)
    8) Cameron denies that Sun has problem with fixing bugs etc.
    9) Cameron claims falsely that Microsoft has become i big bureaucracy where people are feeling uncomfortable
    10) All tests, studies and benchmarks are bough by Microsoft according to Cameron.

    *http://www.theserverside.com/news/thread.tss?thread_id=27341#130452

    And so on and so on, I could go on forever. What shall I say?

    Why do you associate yourself with Cameron Peter?

    Regards
    Rolf Tollerud
  42. I gladly put my foot in my mouth[ Go to top ]

    What shall I say?Why do you associate yourself with Cameron Peter?
    Regards
    Rolf Tollerud

    it has nothing to do with cameron. I'm just a busy body who feels like reminding you the best defense/offense is to stick to the facts. Selectively quoting cameron out of context hardley proves anything to me.

    every technique has a weakness. many people have tried to point out the limits of each technique. even though the debate gets heated, the point is to understand the limitations and learn where each method breaks. Of course we all make mistakes and have to bash our heads to learn the hard lessons, but calling cameron names doesn't really accomplish anything. if you believe that distribute maps is bad, then prove it technically.

    for your information, the domain of distributed maps has been studied for close to 2 decades now. Stanford, MIT, Caltech and several other univerisities have studied it in depth. There are numerous papers on ACMQueue on this topic. From my limited knowledge of this field, the claims cameron makes are in line with the existing research I've read.

    enjoy

    peter
  43. observations on the human nature[ Go to top ]

    For many years ago I used to play Tournament Bridge, and there, as is my wont I made a couple of anthropological observations. You are of course familiar to the form of bridge; 2 players play against 2 players. Many pairs have played together for years.

    I further noticed that those pairs often consist of archetypical roles (you have read your Jung I suppose?) it is the "Teacher", and the "Pupil". No matter that the balance of competence may have changed considerably in the years since they began playing they always keep to the aboriginal setup.

    So how does it work? The "Teacher" pontificates at length before every deal, what the "Student" should do and what he should not do. And then comes the deal. Let us say that the "Teacher" just for some minutes ago has shared with the "Student" some important facts of life that he wants to hammer in, and the result of following the advice is catastrophe, utterly destruction. They are out of the competition.

    Now here comes the interesting. Do you think that the "Student" now revolt and protest vehemently? No. Nobody says anything. The matter is not even broached. The "Student" meekly accepts that this occasion was an "exception", that goes without saying. Funny isn't it? What does it learn us of the human nature?

    That is the explanation to how Cameron can go on year after year saying things that is just plain horrible wrong without anybody contracting him or protesting.

    Regards
    Rolf Tollerud
  44. For many years ago I used to play Tournament Bridge...
    <snip>
    Wrong address, you probably meant to send this to your therapist.

    --
    Cedric
  45. For many years ago I used to play Tournament Bridge, and there, as is my wont I made a couple of anthropological observations. You are of course familiar to the form of bridge; 2 players play against 2 players. Many pairs have played together for years.I further noticed that those pairs often consist of archetypical roles (you have read your Jung I suppose?) it is the "Teacher", and the "Pupil". No matter that the balance of competence may have changed considerably in the years since they began playing they always keep to the aboriginal setup.So how does it work? The "Teacher" pontificates at length before every deal, what the "Student" should do and what he should not do. And then comes the deal. Let us say that the "Teacher" just for some minutes ago has shared with the "Student" some important facts of life that he wants to hammer in, and the result of following the advice is catastrophe, utterly destruction. They are out of the competition.Now here comes the interesting. Do you think that the "Student" now revolt and protest vehemently? No. Nobody says anything. The matter is not even broached. The "Student" meekly accepts that this occasion was an "exception", that goes without saying. Funny isn't it? What does it learn us of the human nature?That is the explanation to how Cameron can go on year after year saying things that is just plain horrible wrong without anybody contracting him or protesting.
    Regards
    Rolf Tollerud

    My area of focus is rather different than Cameron, so I'd have to switch to become a pupil. Though honestly, I'm more interested in Rule engines, compilers, pattern matching algorithms and relational theory. I've researched distributed maps in an attempt to find solutions to challenging scalability problems. You know, those boring problems I keep ranting on about. The ones with real-time requirements and huge databases with millions and hundreds of millions of rows of data.

    If you understood these domains, you'd realize how different the problem domain is. The ACM really does have some good articles with solid scientific research with proper controls. Not those impractical theoriticians you keep bashing. There are plenty of impractical people in academia and in the corporate world. There are also some really bright people who take time to do things thoroughly.

    I would challenge you to use the scientific method to prove Cameron's claims are wrong. I'll gladly admit I am wrong, when you provide repeatable, reliable proof. by the way, you can download a trial version of coherence and easily prove it doesn't work. Not that i think you can prove coherence is wrong or bad, since it's clear tangosol has taken great care to produce a solid product.

    peter
  46. traumatic experience[ Go to top ]

    Peter,

    I agree to that you are probably good in your work. But it is not my field - I am making applications, killing client apps that take the breath out of my customers. You have no interest in my work, I have no interest in yours. It is as it is.

    That has nothing to do with the ten very important areas where I showed Cameron to be wrong, and mark my words, I was just varming up. Cameron has always been wrong. On the other hand, that doesn't make his product a bad product. Kind of wishful thinking I suppose.

    But why are you pestering me? If you think you can impress me think again! Here is another little anecdote clartfy the situaion:

    (from true life)

    One young journalist (his name was Oscar Hedlund actually) was going to interview Andre Segovia, and he was so nervous that he had has spent weeks thinking out the questions. Then he meets Segovia and the first thing he asks is, "What are your opinions of the medieval tertiary tuning?"

    Segovia looks at him, takes the guitar, and handles it over to him and says, "please play a little while I am trying to find out what you were saying".

    Needless to say, after that mr Hedlund always kept a low profile the rest of his life.

    Regards
    Rolf (Segovia) Tollerud
  47. Impress you?[ Go to top ]

    Peter,
    I agree to that you are probably good in your work. But it is not my field - I am making applications, killing client apps that take the breath out of my customers. You have no interest in my work, I have no interest in yours. It is as it is.
    That has nothing to do with the ten very important areas where I showed Cameron to be wrong, and mark my words, I was just varming up. Cameron has always been wrong. On the other hand, that doesn't make his product a bad product. Kind of wishful thinking I suppose.
    But why are you pestering me? If you think you can impress me think again! Here is another little anecdote clartfy the situaion:
    (from true life)One young journalist (his name was Oscar Hedlund actually) was going to interview Andre Segovia, and he was so nervous that he had has spent weeks thinking out the questions. Then he meets Segovia and the first thing he asks is, "What are your opinions of the medieval tertiary tuning?" Segovia looks at him, takes the guitar, and handles it over to him and says, "please play a little while I am trying to find out what you were saying". Needless to say, after that mr Hedlund always kept a low profile the rest of his life.
    Regards

    Rolf (Segovia) Tollerud

    I never claimed to be good either, nor do I care to make any such claims. Honestly, everyone makes mistakes and everyone is wrong a couple of times a day. I don't see the point in telling someone they are wrong. If an individual is reasonable, they'll see they are wrong. Someone unreasonable won't admit they are wrong even in the face of strong evidence.

    I don't see how Segovia has anything to do with technology or the discussion at hand. I pester you in the naive hope that one day you'll answer questions with solid facts. Plus, I enjoy playing devil's advocate.

    I've been wrong plenty of times, but every now and then I get something right. Just because a person makes mistakes, doesn't mean someone "has always been wrong." That's a rather bi-polar view on life. Nothing is purely right/wrong, or black/white. How about trying to stay on topic and stick to technical facts?

    peter
  48. Peter,

    For the last time. I know enough technical facts to do my work and don't care about the rest. Especially as it is highly controversial what is facts or not. For example, this forum can not even agree upon the color of an orange!

    This whole debate started because Cedric said that my statements for the D language "were opinions only". Then I just meekly pointed out that my opinions in the past had turned out pretty accurate. My background makes me better suited to look into the mist of the future. That is my claim and I have proven it.

    Regards
    Rolf Tollerud
  49. Peter, For the last time. I know enough technical facts to do my work and don't care about the rest. Especially as it is highly controversial what is facts or not. For example, this forum can not even agree upon the color of an orange!This whole debate started because Cedric said that my statements for the D language "were opinions only". Then I just meekly pointed out that my opinions in the past had turned out pretty accurate. My background makes me better suited to look into the mist of the future. That is my claim and I have proven it.RegardsRolf Tollerud

    May I meekly point out that a lot of your opinions have been patent nonsense. Who can forget the now-classic "AutoCAD is OOP's ancestral beginning" and wonderfuly self-contradictory posts in which you berate ORM then praise Ruby on Rails. (interestingly and ironically, this even makes your statement "my opinions in the past had turned out pretty accurate" inaccurate). Before you start to look at the future, you should work harder at being able to look at the present and the past.
  50. more nonsense from Steve[ Go to top ]

    You definitely have spent too much time with the wrong sort of people Steve. You decide for yourself on the flimsiest grounds that you have won a discussion and gleefully boast of it everywhere.

    Applications suited to OO according to Alexander Jerusalem:
    "For a windowing system or a flight simulator or a computer game or a graphics package, encapsulation is a desirable means for managing dependencies. These applications own their data, but business applications don't"

    But isn’t that obvious, even for a 7 year old? So you think that the stupid getters/setters classes of ORM qualify as OO? Please don't make me laugh.

    And you have not produced any evidence that eBay uses EJB. I maintain that eBay is one of the costliest projects ever in computer business, it certainly is one of the biggest mistakes of all time. MS could probably have built the whole system in 3-4 months.
  51. more nonsense from Steve[ Go to top ]

    Applications suited to OO according to Alexander Jerusalem:
    "For a windowing system or a flight simulator or a computer game or a graphics package, encapsulation is a desirable means for managing dependencies. These applications own their data, but business applications don't"
    Shouldn't and don't are two different things. Data without an application is useless and an application without data is useless.

    The purpose of an ORM is NOT to separate the application from the data but to separate the persistance mechanism from the domain logic.
    So you think that the stupid getters/setters classes of ORM qualify as OO?
    I highly doubt Steve thinks accessors make an object OO. He is an ex-Smalltalker, after all. He might even think it makes objects less OO (but I wouldn't want to say what other people think - THAT is your job).

    MS could probably have built the whole system in 3-4 months.
    Riiiight. But how could they? They have no Java (or crossplatform) expertise.
  52. more nonsense from Steve[ Go to top ]

    Applications suited to OO according to Alexander Jerusalem:
    "For a windowing system or a flight simulator or a computer game or a graphics package, encapsulation is a desirable means for managing dependencies. These applications own their data, but business applications don't"
    Shouldn't and don't are two different things. Data without an application is useless and an application without data is useless.

    Indeed. And, of course, this has absolutely no relevance to AutoCADs declared status as the 'ancestral beginning of OOP'.
    The purpose of an ORM is NOT to separate the application from the data but to separate the persistance mechanism from the domain logic.

    Very well put!
    So you think that the stupid getters/setters classes of ORM qualify as OO?
    I highly doubt Steve thinks accessors make an object OO. He is an ex-Smalltalker, after all. He might even think it makes objects less OO (but I wouldn't want to say what other people think - THAT is your job).

    Actually, I don't mind them that much. As long as I can treat them as properties and have them managed by an IDE.

    Oh, and for Rolf: Search the web for EBay's presentation at JavaOne 2003 - proof of EJB use. Assuming you don't trust TSS to publish the truth in their article about EBay.
  53. Peter, For the last time. I know enough technical facts to do my work and don't care about the rest. Especially as it is highly controversial what is facts or not. For example, this forum can not even agree upon the color of an orange!This whole debate started because Cedric said that my statements for the D language "were opinions only". Then I just meekly pointed out that my opinions in the past had turned out pretty accurate. My background makes me better suited to look into the mist of the future. That is my claim and I have proven it.
    Regards
    Rolf Tollerud

    I don't care about what you do at work. I only care about how D achieves similar features/functionality as design time API. To do that, all you need to do is talk about how design time API affects language design, just in time optimization and what those trade-offs are. You have 20 years of experience, so I would think that is shouldn't be so hard to do.

    peter
  54. I'll get things started for you[ Go to top ]

    If you compare anonymous classes in .NET 2.0, how dose that differ to design time api.
    http://msdn.microsoft.com/msdnmag/issues/04/05/c20/default.aspx

    Since both anonymous classes and design time API are attempting to address runtime behavior, what are the trade offs of having to discover the metadata of a class.

    peter
  55. Peter,

    You are an idiot. Do you think that we, the professional application authors use a visual design environment? If you knew something about the .NET world you would know that Microsoft’s obsession with design-time elements - to drag and drop components onto the screen is one of the things that they are most criticized most for. You never cease to amaze me. This whole thread, whatever the Java should copy the most stupid/insignificant beginner practice from MS is patetique.
  56. P. S.[ Go to top ]

    Besides all my work is solely in D nowadays.
  57. P. S.[ Go to top ]

    Besides all my work is solely in D nowadays.
    And here you've been telling us that Microsoft is always best.
  58. Peter, You are an idiot. Do you think that we, the professional application authors use a visual design environment? If you knew something about the .NET world you would know that Microsoft’s obsession with design-time elements - to drag and drop components onto the screen is one of the things that they are most criticized most for. You never cease to amaze me. This whole thread, whatever the Java should copy the most stupid/insignificant beginner practice from MS is patetique.

    Dude,
    What is your background? I've never heard of you, yet you speak as if you inventer the computer. You act like an academia who's thesis got shot down, and no you're pissed off because you don't have a site called 'therolfside.com'

    I guess my point is: Put up or shut up. Let's see a resume. Who's using your software now. Feel free to say 'Why does it matter what my background is?' - which will just tell everyone you have no clue what you're talking about.
  59. Have you ever wondered every time you see a very bad movie why it wasn't discovered during the recording of the film? It never is you know, instead they say "everything was so nice, the main actor, the main actress, the director, the producer, everybody was so nice! The whole atmosphere during the shooting was so nice! That’s the way it goes.

    Therefore it is important to have some that strike another chord. You are very new here, only since March 29 and only been participating in two threads. Yet you think that you know so much!

    For your information there is a button that says "Mark as noisy". Has it ever got through to your *** brain that there is a reason that I am not marked as noisy?

    You are a newbie. Please be quite and speak only when you are spoken to. You are allowed to fetch coffee.

    Regards
    Rolf Tollerud
  60. Like I said...[ Go to top ]

    Like I said:
    Feel free to say 'Why does it matter what my background is?' - which will just tell everyone you have no clue what you're talking about.
  61. You are a newbie.
    Well, I'm not, and I'm asking you the exact same thing Michael is asking you.

    Show your resume, describe your work, show some code (didn't you say you were working exclusively in D? I would love to see some D in action... seriously!).

    Do you want to gain some respect? You know what to do.

    But I'm betting you will say that everything you write is proprietary and cannot be disclosed and that we will still have absolutely no evidence that you have any technical knowledge at all.

    Prove me wrong.

    --
    Cedric
  62. Rolf, I have also been developing software for more than 20 years now. The point is that I to spend time surfing TSS and other forums and do not generally feel the need to post as quite often I have an open mind where new technology is concerned or where new languages (pros and cons) crop up. When I scan these threads other people have often covered my point of view so responding would often result in repeated content thus obfuscating the discussion. I also suspect that there are others out there like me. This does not mean that we do not “know our stuff” or that when we do have a point of view that may not have been stated that we should not post. Some of us I expect get real value from the views (sometimes divided) of other people in this community and it is sad when a discussion degenerates into one-up-man-ship or "I bet I can piss higher than you" competitions. And within this thread I now see that most of the posts are not even related to the original topic, remind me now, what was it.... As with 'Michael Finger' you could also say that I was new here, and you would probably be right. But, I also do know my stuff (design and development of massively distributed clustered and load balanced J2EE applications for the financial and telco markets here in Europe) I have been developing J2EE applications since 1997 when it first hit the market with 10 years of C++ prior to that as well as C, fortran, PL1, COBAL, Pascal and basic before that. So your comment to the previous poster has been so broad that you have trawled many others into your .NET (pun intended) and I for one feel that your scope of expertise (whatever that may be on top of front end development) has no bearing on mine or others here and therefore I do not see that you have the necessary skills to rubbish our views or backgrounds just on the number of posts made by an individual. So, in the future it would be appreciated if you would target your responses to the person or persons it is intended for and not the community in general. I personally find no problem with people changing their views over time as technologies come and go. For instance EJB has been mentioned here as a great technology (and was over used) and then rubbished by some of the same people. Things change, better ways of doing things turn up, research gets better and new frameworks emerge. EJB still has its place in J2EE as distributed XA transactions across JVM’s are hard/impossible (currently) to achieve without them (unless you want to get down and dirty with the underlying CORBA code). This does not mean that EJB is a good technology choice in the majority of situations (I would even cut this down to just SLSB facades) but it still is the only choice in a very few situations (please somebody prove me wrong because I have 1 SLSB in my current project that I would dearly love to get rid of). My point being here is that some technologies start off with promise and go some way to solving certain issues and then something new comes along that everybody agrees is a better way (in 99.999% of cases) to do it (this may well happen with D or it may well go down the proverbial pan like so many other once promising languages). Will this current JSR cut the mustard at the end of the day? Only time will tell, but it is obvious that enough people think it will to give it a try, lets see what happens, lets have discussions on the pro’s and con’s and lets also try to stop getting personal if somebody does not like your view. Regards Steven McArdle
  63. In other words[ Go to top ]

    Peter, You are an idiot. Do you think that we, the professional application authors use a visual design environment? If you knew something about the .NET world you would know that Microsoft’s obsession with design-time elements - to drag and drop components onto the screen is one of the things that they are most criticized most for. You never cease to amaze me. This whole thread, whatever the Java should copy the most stupid/insignificant beginner practice from MS is patetique.

    So you're saying you don't know the implication of runtime/design-time techniques. Like for example the article referenced in this thread talks about the comparison between BeanInfo and DesignInfo. Here's a paragraph from the article:
    The new proposed class, DesignInfo, defines dynamic behavior information about a bean instance. Because DesignInfo is discovered at runtime, or design time, existing beans require no changes to benefit from this new functionality.
    I haven't read the proposed spec, but the concept of dynamic class information is very interesting. The design time aspect and how they affect IDE is interesting too. Since you work on rich clients, I would think this topic would be interesting to you. Here's another paragraph from the article:
    The current JavaBeans specifications define a few static ways to relate beans to one another. JSR 273 offers more dynamic abilities to link beans via the linkBean() method on DesignInfo. That will allow a bean to dynamically interact with mouse events, or respond to drag and drop in a customized way.
    In the past you talked about executing DDL in Sql to modify tables. Based on that, I would have thought the capabilities proposed in JSR 273 would be useful to you. I would say any kind of UI that provides modeling capabilities might find the new JSR interesting. Of course i could be totally mis-reading the article and the intent of the JSR, since I haven't read the spec.

    enjoy

    peter
  64. Peter,

    Please take the guitar and play while I wondering what you mean!

    Certainly you know that I do not use an IDE? A syntax highlighted text editor is all you need, and in case of .NET, the standalone debugger.

    In the past you talked about executing DDL in Sql to modify tables

    After thought about it a little while I take it that you mean the old problem when columns are added/deleted during the development or even after.

    Naturally you make your CRUD code 100% generic, if you need to know how to do that with complicated form design involving 4-5 tables o the screen (or more) put me an note, and I will explain. BTW, I never find out how to do that with ORM, that is one of reasons why I don't like ORM.

    To be explicit:
    Adding or deleting 4-5 columns in the database should not require any modifications of your source code.

    Regards
    Rolf Tollerud
  65. To be explicit:Adding or deleting 4-5 columns in the database should not require any modifications of your source code.

    Not having to change code when you add columns in the database is never a problem. But when you delete columns? I'm assuming that you are using at least some of the columns in the database somewhere in your code. How is the source code going to survive removal of these? The only way CRUD code can be that generic is if it doesn't actually explicitly refer to any columns at all... unless I am missing something.
  66. Not only that but modifications od the names of Tables and Columns should not affect the application either. And keep as close as possible to SQL ANSI 92 so changing the database cause as little problems as possible. At last, finally and very important: always store your dates in string format, 'YYYY-MM-DD'.
  67. Not only that but modifications od the names of Tables and Columns should not affect the application either. And keep as close as possible to SQL ANSI 92 so changing the database cause as little problems as possible. At last, finally and very important: always store your dates in string format, 'YYYY-MM-DD'.

    Supposing I have an application that produces a report - say a Tax calculation. In order to do that I refer to a column that contains the value to be taxed - say an 'amount' column. How is the application going to survive removal of the 'amount' column from the database?
  68. There is cultural differences, which must be the reason we don't understand each other.

    We (the other guys) from the client side see things a little different. In a typical Web-app everything is fused together in an unfamiliar way for us Rich-Clients guys. We usually have clear separation of the application in these three parts:

    1) The interactive user interface where the customer spent 90% of his time, doing his work.
    2) The reports, that is not interactive (instant).
    3) The report generator where you design your reports.

    The thing is that changes in the database may of course break your reports, but that is not a big deal because the user has control over the report-design and can mend the report herself.

    It is only about KISS and about not overcomplicate things. For example, if Vic Cekvenich should show up his code here in TSS I am sure it would not impress anybody, still his solutions work! That is all the difference.

    Likewise, I am sure my code would not impress anybody either. (I remember how hard I was attacked and ridiculed at one time before when I advocated store of dates in 'YYYY-MM-DD' string format. Nevertheless. I am adamant)

    Regards
    Rolf Tollerud
  69. In a typical Web-app everything is fused together in an unfamiliar way for us Rich-Clients guys. We usually have clear separation of the application in these three parts:
    1) The interactive user interface where the customer spent 90% of his time, doing his work.
    2) The reports, that is not interactive (instant).
    3) The report generator where you design your reports.
    The thing is that changes in the database may of course break your reports, but that is not a big deal because the user has control over the report-design and can mend the report herself.
    Regards
    Rolf Tollerud

    Ok, having worked on reporting for compliance applications, I'm gonna say you've only scratched the surface. Writing an application to do offline reports is completely different than generating arbitrary adhoc reports for given time interval in real-time. Believe it or not, there are a ton of financial applications that generate real-time reports ranging around 100-200mb. that's not even the extreme either.

    Even in the simple case of letting the end user "design and mend the report", it is still a good idea to have a clean way of defining that report. Unless you're talking about static reports and not adhoc reports.

    peter
  70. This isn't KISS[ Go to top ]

    There is cultural differences, which must be the reason we don't understand each other.We (the other guys) from the client side see things a little different. In a typical Web-app everything is fused together in an unfamiliar way for us Rich-Clients guys.The thing is that changes in the database may of course break your reports, but that is not a big deal because the user has control over the report-design and can mend the report herself.It is only about KISS and about not overcomplicate things.

    This just won't work for almost all applications, rich client or not. Where I work the user doesn't get to change the much of the application, as the application contains critical financial code that has to be audited and checked - we don't want the casual user to be changing the way Tax is calculated! The reports are critical and have to be approved and tested by the financial experts and can't be arbitrarily modified.

    This is why ORM is good - the mapping of the database to the code is simple, explicit, and in one place. With strongly typed languages the relationship of the code to the database can be largely tested at compile time and, where appropriate, changes can be made with refactoring tools.

    You don't want a report or large financial calculation to break a day or two before a deadline because someone has changed the database!
    if Vic Cekvenich should show up his code here in TSS I am sure it would not impress anybody, still his solutions work! That is all the difference.Likewise, I am sure my code would not impress anybody either.

    This is simply bad coding, nothing else. In almost all development situations you personally don't own the code. It should be able to be handed over to other developers at short notice. Any other approach is simply unprofessional - sorry.
  71. to kick and scream[ Go to top ]

    Steve,

    I would be happy to answer your post Steve if it was even remotely connected to what we are discussing at the moment. As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?

    To sum up everything in one sentence: Microsoft is better than the opposition.

    That’s it. You can deny it, cry, protest, scream, do what ever you want, but in the end it is the undeniable truth.

    Regards
    Rolf Tollerud
  72. to kick and scream[ Go to top ]

    Steve,I would be happy to answer your post Steve if it was even remotely connected to what we are discussing at the moment.

    You said:
    To be explicit:
    Adding or deleting 4-5 columns in the database should not require any modifications of your source code.

    and
    Not only that but modifications od the names of Tables and Columns should not affect the application either.

    and
    The thing is that changes in the database may of course break your reports, but that is not a big deal

    I gave a specific example of how deleting columns will require modifications of the code, and how breaking reports could cause major problems (when those reports have to be tested, approved and audited and aren't just something hacked together by the end user).

    This was directly replying to what we are discussing at the moment.
    Microsoft is better than the opposition.

    Which - regardless of it's validity or otherwise - has absolutely no relevance.

    Perhaps we might get somewhere if you go along with what others have suggested and (if you are permitted by commercial considerations) give us some idea of what you actually do. This might put your continual suggestion that almost all IT projects can be developed as quick-and-dirty hacks with scripts and simple tools into perspective. You need not be specific - a general idea of what you work on would be interesting.
  73. to kick and scream[ Go to top ]

    To sum up everything in one sentence: Microsoft is better than the opposition.

    To sum up everything in one sentence: Rolf is wrong, so he tries to change the subject to any random thing .. hmm .. say Microsoft?

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  74. to kick and scream[ Go to top ]

    Steve,I would be happy to answer your post Steve if it was even remotely connected to what we are discussing at the moment. As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    To sum up everything in one sentence: Microsoft is better than the opposition.
    That’s it. You can deny it, cry, protest, scream, do what ever you want, but in the end it is the undeniable truth.
    Regards
    Rolf Tollerud

    Again, you go off topic. Steve and I attempted to stay "some what" on topic and stick to technical facts. All you have to do is show how deleting a row doesn't affect your application using your "CRUD" approach. This has nothing to do with you, it has to do with the "CRUD" technique. Alot of people have tried a variety of techniques to shield applications from database changes, so most of us are aware of the most common techniques.

    I'm only guessing here, but I'm assuming when you say delete a row doesn't affect the application, it is because several columns are read to one alias using "as". Regardless of the approach, when a critical column is deleted and replaced with a new table, it's best to update the application. Feel free to provide a concrete example backing your assertion that add/deleting columns from a table doesn't affect the application.

    peter
  75. to kick and scream[ Go to top ]

    Again, you go off topic. Steve and I attempted to stay "some what" on topic and stick to technical facts. All you have to do is show how deleting a row doesn't affect your application using your "CRUD" approach.

    I was trying to stick to technical facts, but I think that was probably a waste of time. A statement that columns can be deleted and never have an effect on code is such patent nonsense it is hardly worth trying to refute. It is like stating that code should no never depend on the structure of data.

    I get the impression that Rolf has a tendency to unintentionally make a sweeping statement, something like: 'deleting columns should not effect any code', instead of something more reasonable like 'code should be written so as to easily cope with changes in database structure'.

    I think the problem is that he then refuses to admit a mistake. Therefore he defends these sweeping statements with ever wilder claims, irrelevant quotes, distractions and insults.

    Whereas the rest of us (at least me) are distinctly fallible, Rolf seems to have to maintain an impression of infallibility for some reason. This would also explain the regular switching back to points of view which have regularly been disproved.

    My conclusion is that arguing with him is fun (at least for me), but won't get anywhere....
  76. agree :)[ Go to top ]

    Again, you go off topic. Steve and I attempted to stay "some what" on topic and stick to technical facts. All you have to do is show how deleting a row doesn't affect your application using your "CRUD" approach.
    I was trying to stick to technical facts, but I think that was probably a waste of time. A statement that columns can be deleted and never have an effect on code is such patent nonsense it is hardly worth trying to refute. It is like stating that code should no never depend on the structure of data.I get the impression that Rolf has a tendency to unintentionally make a sweeping statement, something like: 'deleting columns should not effect any code', instead of something more reasonable like 'code should be written so as to easily cope with changes in database structure'.I think the problem is that he then refuses to admit a mistake. Therefore he defends these sweeping statements with ever wilder claims, irrelevant quotes, distractions and insults.Whereas the rest of us (at least me) are distinctly fallible, Rolf seems to have to maintain an impression of infallibility for some reason. This would also explain the regular switching back to points of view which have regularly been disproved.My conclusion is that arguing with him is fun (at least for me), but won't get anywhere....

    yeah, it definitely goes no where. kind of like waiting for godot (http://www.theatrehistory.com/french/beckett002.html) or six characters in search of an author.

    peter
  77. We're the idiots.[ Go to top ]

    <blockquoteMy conclusion is that arguing with him is fun (at least for me), but won't get anywhere....
    It seems pretty pointless - Rolf won't offer up any facts to back up his arguments. Talking to people that have different views then you is a great way to learn *if* they provide facts and concrete examples to make you broaden your horizons.

    Interacting with Rolf goes like this:

    Rolf - X is way to do it, your way, Y ,is wrong.

    Anyone else - Rolf, I disagree. Y works because of these concrete reasons. Please give me an example of how X would work in these situations.

    Rolf* - You are an idiot. I am really ashamed to have to explain this to you! Let us stop this nonsense and put matters into perspective! Come down to earth.
    *Actual 'rolf-isms'

    We're the idiots for engaging this guy. It's a waste of time (it's really not that entertaining either - you saw the same show in the last thread) until he puts up or shuts up. It's like arguing with a realigious zealot...

    I really don't give a rats ass if .net or java is better - There are people that don't have a choice (Right now, I'm developing for AIX - and the Microsoft choices are - uh - limited.)
  78. We're the idiots.[ Go to top ]

    My conclusion is that arguing with him is fun (at least for me), but won't get anywhere....
    It seems pretty pointless - Rolf won't offer up any facts to back up his arguments.

    On the contrary, Rolf frequently provides a lot of 'facts' to back up his arguments. The problem for him is that not only do these facts usually not back up his arguments, they often end up being good evidence against his point of view. Recent classics included rants against ORM in which he praised Ruby on Rails (which includes the ActiveRecord ... ORM!).
    We're the idiots for engaging this guy. It's a waste of time (it's really not that entertaining either - you saw the same show in the last thread) until he puts up or shuts up. It's like arguing with a realigious zealot...

    My view is that FUD left unchallenged might have some influence. But you are probably right....
  79. theserverside vs theclientside[ Go to top ]

    Well, today it is Monday morning and I apologize for not being able to participate in the charming conversation.

    The reason that I was a little bit irritated at the end (sorry I lost my temper) was that I am tired of this practice of introducing "the world largest financial systems" into every debate.

    Let us first finish that debate once and for all and be done with it. "My opinions of the largest systems in the world, financial or otherwise is what follows",

    either,

    1) Microsoft already has a better price/performance solution or,
    2) Microsoft will have a better price/performance solution in the future.

    Which one it is right now I really don't care because I don't work in that field! Let other worry about it. But if you everytime we are discussing how to build summer-cottages you shall drag "Queen Elisabeth’s summer-cottage" into the discussion that is a little boring, to use an understatement.

    There is an other aspect on the situation, namely: Money.
    People should remember the Fortune 5000 is composed of just about 5000 businesses. In the United States alone There are over 5.5 million small businesses with employees. There's a huge commercial middle ground and it's several orders of magnitude larger than the "enterprise" market.


    Are you not fascinated for example of salesforce.com that with a little program built in a few months practically overnight turned into a multi billion business? Hmm, then US citizens can no longer be Americans, IMO! :)

    And to if it is possible to let the customers add/delete fields that is kind of a moot question, it is common practice in our world - get used to it. For example the above mentioned saleforce.com uses it. It works well enough in practice, it is proven in the field.

    It is always so fun when people argue with reality.

    To sum it up. It is easy to make systems that not allow the user do to anything but in the real world (I almost said jungle :) there is a typical dangerous kind of species that can not be ignored, that is called the "Power-user".

    In confrontation with those types of persons, who usually hava an IQ of 150 or more, you need to come prepared. To have a professional fast program with lots of lots of functionality, hooks and APIs, professional report-generator and hundreds of predone report, etc, etc helps. Otherwise you will have problems.

    Regards
    Rolf Tollerud
    (To Floyd: It is better to rename theserverside.net to theclientside.com(net), to handle "the enterprise client side". There must be a huge need! ;)
  80. theserverside vs theclientside[ Go to top ]

    Another example of a typical Rolf post..
    Are you not fascinated for example of salesforce.com that with a little program built in a few months practically overnight turned into a multi billion business?

    Which, of course, uses almost all of the technologies that Rolf rants about. The 'little program' was developed using Java and Oracle on Solaris (they now use Linux as well).

    Also, typically for Rolf, he provides an example which contradicts not just his technology opinions, but his views. In a recent interview (last month) the CEO of salesforce.com "castigated Microsoft for failing to deliver the next wave of innovation", and said: "The Blackberry, iPod and Salesforce.com exist because Microsoft let us down".

    Yet again, I thank him. I now have even more useful quotes and evidence to promote use of non-MS technologies (where appropriate, of course).
  81. theserverside vs theclientside[ Go to top ]

    1) Now you suddenly dropped the subject of added/delete columns that was the subject of the discussion.

    2) "The Blackberry, iPod and Salesforce.com exist because Microsoft let us down"

    So you bashing MS for letting some innovation still go on outside Microsoft. hi hi

    "The 'little program' was developed using Java and Oracle on Solaris (they now use Linux as well)"

    Since when have I been against Java used in "Vic Cekvenich" style? That is how they do things on salesforce.com

    In fact at salesforce they do almost everything exactly opposite to normal behavior in J2EE circles. They put large resources on the client interface for example, a normal J2EE user form looks like something autogenerated from a tool. And in addition to let users add/delete fields, they also let power-users develop customized extensions. They have a good reporting tool. For just a few posts earlier in the thread you were advocating that the users should not be permitted to do anything, remember?

    Java, Oracle, Linux ok. But everything else is different. Why trying to credit J2EE with something that is almost totally opposite to normal J2EE practice?

    Why don't you see and learn instead? A normal J2EE project can be compared to an elephant in a glass store.
  82. Untrue[ Go to top ]

    That is how they do things on salesforce.comIn fact at salesforce they do almost everything exactly opposite to normal behavior in J2EE circles. They put large resources on the client interface for example, a normal J2EE user form looks like something autogenerated from a tool. And in addition to let users add/delete fields, they also let power-users develop customized extensions. They have a good reporting tool.

    Again, I ask you, where is the proof. I seriously doubt they handle add/delete of columns the way you do it. In other words, they don't do "alter table ...". Most likley, they are using mapping technique to provide custom features. If you knew anything about bandwidth requirements, you'd realize how wrong you are about "they put the large resources on the client". No business would use salesforce.com without strong encryption. Therefore, that means they can't simply load as much as possible in the client. That would kill their bandwidth and it would make their entire platform unstable.

    Instead, they have to be selective about what they send and try to use the bandwidth efficiently. I haven't looked at their reporting features, but I seriously doubt they handle reports the way you do it. Most large applications will utilize a combination of the OLAP features and summary tables to balance performance vs flexibility.

    Feel free to prove me wrong and show how salesforce.com handles custom extensions. My guess is they have their own plugin API and mapping framework to make it simple and manageable.

    peter
  83. what is it that its ntrue[ Go to top ]

    "If you knew anything about bandwidth requirements, you'd realize how wrong you are about "they put the large resources on the client"

    Peter, I know of only one person that misunderstands more that you. But Cameron does it on purpose! ;) Therefore it is difficult to talk to you.

    By "putting the large resources on the client" I mean that they have allocated many development hours to the client, contrary to normal J2EE practice.

    they don't do "alter table ...".

    Why should they have to????

    "I haven't looked at their reporting features, but I seriously doubt they handle reports the way you do it

    On the contrary they do it exactly the way I would have done it.

    "show how salesforce.com handles custom extensions"

    Please I have other things to do in life and there are many ways to do it,
    just read this so you at least have a proof of what they have,
    http://searchcrm.techtarget.com/originalContent/0,289142,sid11_gci901088,00

    On the other hand if you are interested I can show you how we did it in VB many years ago..

    Regards
    Rolf Tollerud
  84. what is it that its ntrue[ Go to top ]

    "If you knew anything about bandwidth requirements, you'd realize how wrong you are about "they put the large resources on the client"Peter, I know of only one person that misunderstands more that you. But Cameron does it on purpose! ;)

    oh come on. so now you're blaming me for your post be unclear. ok, my fault for not reading your thoughts. Oh wait, I can't read people's thoughts, so it's my fault for not being a mind reader.
    Therefore it is difficult to talk to you. By "putting the large resources on the client" I mean that they have allocated many development hours to the client, contrary to normal J2EE practice.they don't do "alter table ...".Why should they have to????"I haven't looked at their reporting features, but I seriously doubt they handle reports the way you do itOn the contrary they do it exactly the way I would have done it.

    and what technique would that be? you've stated in the past you have no experience with OLAP. So either your statement in the past was wrong, or you believe salesforce.com uses pure Sql. which one is it?

    <blocquote>"show how salesforce.com handles custom extensions"Please I have other things to do in life and there are many ways to do it, just read this so you at least have a proof of what they have,http://searchcrm.techtarget.com/originalContent/0,289142,sid11_gci901088,00On the other hand if you are interested I can show you how we did it in VB many years ago..
    Regards
    Rolf Tollerud
    I downloaded their SDK and from their examples, it would appear salesforce.com uses the mapping technique described by myself and others. I'm just guessing based on their sdk, but it would appear they absolutely do not use DDL to provide custom extensions. They use mapping techniques. That makes sense, since it allows them to keep their database model stable and reduce the amount of code changes.

    It would have only taken 10 minutes to download the SDK from salesforce.com and figure out how they do things, yet you do not. Is that because you're afraid of proving yourself wrong?

    peter
  85. OLAP? Now you are coming with "Queen Elisabeth’s Summer-Cottage" again. Before OLAP, you have to have basic SQL reporting. To choose columns, type and locations, simple sums, different design templates, SQL selection, etc.

    So you downloaded their SDK. I say that your explanation is shit.

    It really is difficult to get good help these days isn't it?.
  86. OLAP? Now you are coming with "Queen Elisabeth’s Summer-Cottage" again. Before OLAP, you have to have basic SQL reporting. To choose columns, type and locations, simple sums, different design templates, SQL selection, etc.So you downloaded their SDK. I say that your explanation is shit.It really is difficult to get good help these days isn't it?.

    download their SDK and look at:

    DescribeGlobalResult.java
    DescribeLayoutResult.java

    Given that I don't work for salesforce.com, I admit I could be totally wrong. But from their SDK, one can guess the approach they are using. I think it's already been established you have to know Sql. If you knew anything about OLAP, you'd realize that OLAP provides a way to map table columns to cubes dimensions. go look at the screen shots of analysis service. When a user defines a cube, they map a dimension to the table.column. The next step is the measure is mapped to a function on one or more columns.

    By doing this, it becomes easier to build tool that allow users to define new reports in an adhoc manner. Without doing this, the end user would have to know the table structure and Sql fairly well to generate reports efficiently. That is all basic reporting techniques right. One can even use a trigger + indexed view to create a summary table. But obviously, that slows down the inserts and may not be acceptable for a database with heavy transactions.

    enjoy

    peter
  87. Please I have other things to do in life
    Well, that's the only proof (or truth) we actually need, and none other.
  88. theserverside vs theclientside[ Go to top ]

    So you bashing MS for letting some innovation still go on outside Microsoft. hi hi

    I wasn't bashing anyone. It was the CEO of salesforce.com
    For just a few posts earlier in the thread you were advocating that the users should not be permitted to do anything, remember?

    Nonsense. Please stop misquoting. I said
    Where I work the user doesn't get to change the much of the application, as the application contains critical financial code that has to be audited and checked

    There are parts of the application the user certainly CAN change, and some reports they are permitted to design and modify. However, most of it is complex financial calculations. I assume, Rolf, when you use a bank you would prefer it if some parts of the way the bank operates is sealed off from casual modification by 'users' and approved by accountants, legal experts etc?
    Why trying to credit J2EE with something that is almost totally opposite to normal J2EE practice?

    Because you keep setting up this straw man of what YOU think is 'normal J2EE practice'. There is almost certainly a vast amount of J2EE that runs without EJB or much else - just JSP/Servlet, security etc, running above databases. This is the environment where Spring and Hibernate are widely used.

    Perhaps you have some sort of international developer and project survey that illustrates what 'normal J2EE practice' is?
  89. theserverside vs theclientside[ Go to top ]

    hi hi
    Rolf, I think what you are trying to spell is hehe (laugh), rather than saying hello (hi) to everyone twice in quick succession.
  90. language lesson[ Go to top ]

    Thank you Jack, for reminding me.

    hi hi is a reminiscence from my native language. However he he is definitely not right either, what I mean to convey is a meaning more like snickering, more like immature schoolgirls giggling when small boys make stupid remarks! Perhaps you can help?

    English is a hard language to learn, sigh
  91. IMO an APIs alone never made coding easier or harder. If you want to simplify rich client development (be it on the desktop or the web), start with an infrastructure. Do it right on the architecture level first and then focus on the IDE side of it. Maybe the right decision woudl be to dump Swing and re-build it with another architecture.

    Yes, there is a need for real desktop clients and unfortunatley Swing is all we have in Java for this.

    Frank
  92. Oh really?[ Go to top ]

    Are you not fascinated for example of salesforce.com that with a little program built in a few months practically overnight turned into a multi billion business? Hmm, then US citizens can no longer be Americans, IMO! :)
    And to if it is possible to let the customers add/delete fields that is kind of a moot question, it is common practice in our world - get used to it. For example the above mentioned saleforce.com uses it. It works well enough in practice, it is proven in the field.
    It is always so fun when people argue with reality. To sum it up. It is easy to make systems that not allow the user do to anything but in the real world (I almost said jungle :) there is a typical dangerous kind of species that can not be ignored, that is called the "Power-user". In confrontation with those types of persons, who usually hava an IQ of 150 or more, you need to come prepared. To have a professional fast program with lots of lots of functionality, hooks and APIs, professional report-generator and hundreds of predone report, etc, etc helps. Otherwise you will have problems.
    Regards
    Rolf Tollerud(To Floyd: It is better to rename theserverside.net to theclientside.com(net), to handle "the enterprise client side". There must be a huge need! ;)

    Where did you get the statistic that it took salesforce.com a few months to build their current product? Building a sophisticated CRM takes a lot more than just 3-4 months. Getting the production hardware, finding the service provider, and configuring the servers will take at minimum 4-5 months. That's assuming you already have the application finished and are ready to roll out.

    When salesforce.com first came out, they didn't provide the level of features they have today, and I seriously doubt their architecture has remained the same. I don't work for them, so I don't know exactly how their setup is. I find it hilarious you feel it is acceptable to speak on behalf of Salesforce.com again and again without any concrete information about their architecture and design. I do know of other business like Netledger, which was purchased by Oracle. netLedger used apache, jserv, SSL, and oracle when they first started. The requirements of a business like salesforce.com or netledger have completely different requirements than an internal rich client application.

    Feel free to prove me wrong. Sorry to everyone for being off topic.

    peter
  93. ?[ Go to top ]

    1) How can salesforce possible take offense to all that praise I heap upon them?

    2) How can the Java community take offense to such a fine example of what Java can do?

    3) The only ones that possible can make a sour face is the fake J2EE "scientist".

    (How come the subject of "added/deleted columns suddenly disappeared?)

    "When salesforce.com first came out, they didn't provide the level of features they have today"

    That goes without saying. But the initial program was enough for instant success.

    Regards
    Rolf Tollerud
  94. ?[ Go to top ]

    How come the subject of "added/deleted columns suddenly disappeared?

    Because when I provided evidence to show you were wrong you accused me of wandering off topic.
  95. Add/delete comun example[ Go to top ]

    How come the subject of "added/deleted columns suddenly disappeared?
    Because when I provided evidence to show you were wrong you accused me of wandering off topic.

    Ok, here's a add/delete columns example/question. I think it's pretty simple and a real life example...Let's say I work for a health care insurance company on the claims processing side, and we're getting rid of social security numbers and replacing them with customer id's. The social security number column shows up on 16 screens, but we only want to change them on 14 (2 of the colums are used for fraud detection and tax reporting - so we need social) We're adding a new field to the database and we need to switch the colums out for some of the screens. There is also a data validation we need to do against the data, which we do with a rules engine. The rules live in the database. So I assume we would have to keep alot of display data and screen->data mappings in the database. We're also not going to change any of this stuff without a full QA test cycle. So, it seems to me like I would be changing the 'source code' to add and delete a column? And by the time I've changed mappings, rules, and QA, I've spent as much time as I would changing some server side code?
  96. The answer is of course, that it is not allowed to change/delete important fields. User defined fields do no belongs in that category but is still very useful. You could have reasoned out that by yourself couldn't you?

    Since the good members of TSS can not take this discussion seriously I think we better set a point for it right now. Or what do you think? At least I put a stop to answering anything that is not at least moderately intelligent!

    It is the 9th of may today!
    I can inform you that the "Yann Caroff statistic" went up 3% this month instead of the usual 1%. Is it a trend? We see next month!

    Last month 47%
    This month 50%
    Regards
    Rolf Tollerud
  97. why do you not ask salesforce?[ Go to top ]

    The answer is of course, that it is not allowed to change/delete important fields. User defined fields do no belongs in that category but is still very useful. You could have reasoned out that by yourself couldn't you?

    This was in direct contrast to what you said:
    Adding or deleting 4-5 columns in the database should not require any modifications of your source code.

    You could, of course, have said '4-5 user defined fields', but you didn't. As usual you generalised.
    I can inform you that the "Yann Caroff statistic" went up 3% this month instead of the usual 1%. Is it a trend? We see next month!Last month 47%This month 50%

    And the latest "Steve Statistic" - jobs from monster.fr - shows VB.NET fading fast:

    VB.NET 32
    COBOL 52

    See! I can quote statistics that don't prove anything too!
  98. are we suppose to read your mind[ Go to top ]

    The answer is of course, that it is not allowed to change/delete important fields. User defined fields do no belongs in that category but is still very useful. You could have reasoned out that by yourself couldn't you?
    Regards

    Rolf Tollerud

    Again, you blame everyone else when you over generalize. Are we suppose to read your mind and figure out that when you said add/delete columns shouldn't require changes you meant "user specific extensions."

    I still find the statement above overly general. I can think of a scenario where the above statement starts to fall apart. say that i provide an accounting service over the internet. The business needs to define custom attributes, which are critical to tracking inventory. Rather than speak abstractly, i'll use a real example. say I generate an internal unique ID for each product and map it to the manufacturer's SKU id. Let's call this internal_prod_id.

    If I use that id to track all inbound and outboud items at the factory, that id is critical. Now, lets say my sales guys want to create special discounts and assign a sub id for those items. Each sales person could define a sub id, which is the combination of the internal_prod_id and the sub id defined by the sales person.

    now, lets say each day we want to generate a sales report for that sub id, so that we can track how effective the sale was. In this case, not only is the internal_prod_id critical, but many other features depend on it. Obviously, you could define the internal_prod_id as critical, but you'd then have to keep a record of which custom attributes are critical.

    take this a step further. Say that I want to generate a report every week by internal_prod_id and take into account any specials. that means, my report has to include all sub id in the aggregations totals. Now, if I want to slice and dice the sales patterns of every product by sub id, the application has to be able to map those relationships.

    In cases like these, some form of visual modeling would be nice. having a clean way of relating the visual model with the underlying database and OLAP structure is also needed to make these types of applications easier to build. I've built frameworks to manage these types of relationships, because without it, the user wouldn't be able to create adhoc reports.

    peter
  99. clarification[ Go to top ]

    Just to be safe, when I said "user" in the previous post, I am referring to an end user who does not know Sql, MDX, Databases or care to know. The end user might be a sales person, secretary or marketing person.

    peter
  100. a paus please[ Go to top ]

    I'll be back
  101. be patient[ Go to top ]

    Ok Peter I follow you but in the case you mentioned I would not create a sub_id but a category in an entirely different table. We have to make some clarifications here. There is much more with the rich-clients that we haven't speaked about yet!

    1) A generic form is advantageous even for your own development and maintenance of the product. The user add/delete column feature is an "spin-of" effect that you may or may not choose to use.

    2) One more important features of the rich-client is the category setter. The user should be able to set an unlimited number of categories on any table of her choice, and later be able to search for these categories with a simple point and click procedure. The master group/subcategories table is set by special-permit power users.

    3) Another helpful thingy is the ad-hoc search editor, where users can build up SQL-queries. They only see a = 'b' c = 'd' and so on, the intricacies of SQL is hidden for them. (But again - special-permit power users can choose to see the generated SQL and possible modify it).

    So we see we have barely begun to scratch the surface of the rich-client!

    Regarding your example with OLAP I don't expect the salesforce to have OLAP, wouldn't that permit one user to hi-jack the whole server? But maybe there is some way.

    Regards
    Rolf Tollerud
  102. be patient[ Go to top ]

    Ok Peter I follow you but in the case you mentioned I would not create a sub_id but a category in an entirely different table. We have to make some clarifications here. There is much more with the rich-clients that we haven't speaked about yet!
    1) A generic form is advantageous even for your own development and maintenance of the product. The user add/delete column feature is an "spin-of" effect that you may or may not choose to use.
    2) One more important features of the rich-client is the category setter. The user should be able to set an unlimited number of categories on any table of her choice, and later be able to search for these categories with a simple point and click procedure. The master group/subcategories table is set by special-permit power users.
    3) Another helpful thingy is the ad-hoc search editor, where users can build up SQL-queries. They only see a = 'b' c = 'd' and so on, the intricacies of SQL is hidden for them. (But again - special-permit power users can choose to see the generated SQL and possible modify it).

    So we see we have barely begun to scratch the surface of the rich-client!Regarding your example with OLAP I don't expect the salesforce to have OLAP, wouldn't that permit one user to hi-jack the whole server? But maybe there is some way.
    Regards

    Rolf Tollerud

    first off, thanks for finally responding with specifics. I agree with you, it's just the start. In terms of OLAP and users "hi-jacking" the server. The is one of the benefits of OLAP. Unlike summary tables (ie triggers + indexed views) is that you're not precalculating. Older OLAP products use pre-calculated approach, which resulted in dramatic increase in storage requirements. In essence, the pre-calculated approach is close to combinatorial and requires careful consideration to avoid exploding the database size.

    Since most OLAP servers use bitmap indexes, it actually only joins on the indexes and not the values. Therefore, the cost of running adhoc queries need not "hi-jack" the server is most cases. Where it would hi-jack the servers is using Relational-OLAP, which does a direct query against the primary database. In those cases, analysis server simply issues optimized queries against Sql Server. Another area where OLAP is inappropriate is getting the distinct count when the data changes rapidly. In those cases, it is much better to query against sql server with "select distinct col from blah....".

    the downside with using OLAP is that any changes to the database table requires updating the mapping of the cube to the tables. Therefore, a mapping approach for custom attributes is the preferred approach. I've ranted on enough. There's far more to reporting that just running sql queries or rich clients.

    peter
  103. be patient[ Go to top ]

    here's far more to reporting that just running sql queries or rich clients.peter

    +1

    Which is why, in the general case, you can't just arbitrarily add or delete columns without major code changes.
  104. he is completly lost[ Go to top ]

    There's far more to reporting that just running sql queries or rich clients

    Sometimes you have to read between the lines. I sense a discrepancy, an anomaly in the balance of the universe. You talk as if reporting is all there is to it. On the contrary, the typical salesforce user only draw out some reports at the end of the day, or maybe even only once a week.

    Still, maybe she has been sitting with the system the whole day every day in the week! You have no idea of what a salesperson do, do you?

    Sometimes I wonder if it is ever possible to get therserverside and theclientside to understand each other, it is like "Men from Mars, Women from Venus".

    We are the men, of course.

    Regards
    Rolf Tollerud
  105. first hand experience[ Go to top ]

    There's far more to reporting that just running sql queries or rich clients
    Sometimes you have to read between the lines. I sense a discrepancy, an anomaly in the balance of the universe. You talk as if reporting is all there is to it. On the contrary, the typical salesforce user only draw out some reports at the end of the day, or maybe even only once a week.Still, maybe she has been sitting with the system the whole day every day in the week! You have no idea of what a salesperson do, do you?Sometimes I wonder if it is ever possible to get therserverside and theclientside to understand each other, it is like "Men from Mars, Women from Venus". We are the men, of course.
    Regards
    Rolf Tollerud

    That may be true in your experience, but I've seen cases where adhoc queries are performed in real-time and several times a day. The reporting world is rather diverse and there are a lot of companies out there trying to improve the ease of generating reports. OLAP has been around for over 20 years, but only recently has more companies start using it. Microsoft deserves credit for helping to bring OLAP to a wider audience. Prior to Analysis service, which was written by another company, OLAP was expensive and only a few big companies used it. Today, more and more businesses are using it to make reporting easier and the requirements for adhoc reports is increasing.

    peter
  106. he is completly lost[ Go to top ]

    We are the men, of course.RegardsRolf Tollerud
    However he he is definitely not right either, what I mean to convey is a meaning more like snickering, more like immature schoolgirls giggling

    Those voices in your head are not healthy for you, Rolf - or what's the name of the little girl I *may* be talking to right now..?
  107. he is completly lost[ Go to top ]

    ...it is like "Men from Mars, Women from Venus".

    Looks like Rolf is from Pluto....

    Regards
  108. report from pluto[ Go to top ]

    Celebrating this months jump from 1% to 3% in the Yann Caroff index (= the rate C# is overtaking Java) I give you another survey that further confirms,

    Computerworld Development Survey gives nod to C#
    It beat out Java as the preferred programming language in a survey of developers
        
    http://www.computerworld.com/developmenttopics/development/story/0,10801,100542,00.html

    Regards
    Rolf Tollerud
    (I love to say: "I told you so")
  109. But what about D?[ Go to top ]

    Celebrating this months jump from 1% to 3% in the Yann Caroff index (= the rate C# is overtaking Java) I give you another survey that further confirms,Computerworld Development Survey gives nod to C# It beat out Java as the preferred programming language in a survey of developers &nbsp;&nbsp;&nbsp;&nbsp;http://www.computerworld.com/developmenttopics/development/story/0,10801,100542,00.htmlRegardsRolf Tollerud(I love to say: "I told you so")

    what happened to your new favorite language? What's the job market like for D?

    just kidding

    peter
  110. report from pluto[ Go to top ]

    I give you another survey that further confirms,Computerworld Development Survey gives nod to C# It beat out Java as the preferred programming language in a survey of developers &nbsp;&nbsp;&nbsp;&nbsp;http://www.computerworld.com/developmenttopics/development/story/0,10801,100542,00.htmlRegardsRolf Tollerud(I love to say: "I told you so")

    "A total of 966 qualified completed surveys were received"
    Excellent sampling. Another "F" for Rolf on interpreting surveys and statistics.
  111. Pluto? More like Uranus[ Go to top ]

    That is what my title should have been. :)
  112. report from pluto[ Go to top ]

    I give you another survey that further confirms,Computerworld Development Survey gives nod to C# It beat out Java as the preferred programming language in a survey of developers &amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;http://www.computerworld.com/developmenttopics/development/story/0,10801,100542,00.htmlRegardsRolf Tollerud(I love to say: "I told you so")
    "A total of 966 qualified completed surveys were received"Excellent sampling. Another "F" for Rolf on interpreting surveys and statistics.

    Absolutely. That survey is hilarious. Any survey that lists .Net vs. Unix/Linux as alternative 'Framework/API's shows little or no understanding of what developers are doing. I mean, who on Earth did they find who said 'I use the Linux API'? The report is also bewilderingly contradictory. C# was reported as apparently having being used by more developers than Java, but that has no indication of how much it is used. The same report says that 37% used mostly Java, while 26% said they used mostly .Net, with 23% using both. This means that Java is the most used language!

    Yet again, continuing with his usual record, Rolf provides evidence that supports the contrary view. The difference is between answer to the questions "have you even used this language?" (C# wins) and "what do you mostly use?" (Java wins).
  113. resistance is futile[ Go to top ]

    "This means that Java is the most used language!"

    I never have said anything else. What I have said is,

    1) .NET is most used in new projects and
    2) .NET and C# is gaining on Java by ca 1% every month.

    "The same report says that 37% used mostly Java, while 26% said they used mostly .Net, with 23% using both"

    That dovetails nicely with other reports. But that difference is fast passing away alas.

    I predict that .NET will have overtaken Java before the end of this year. By that I mean that the search: c# or .net will yield more jobs than the search java or j2ee at www.it.jobserve.com before 2005-12-31.

    Regards
    Rolf Tollerud
  114. No love for D[ Go to top ]

    But what about D. have you already forsaken D and gone back to C#.

    this message will self destruct in five seconds

    peter
  115. Peter is a careful person[ Go to top ]

    Why D is under the radar is a constant amazement to me. Especially in the gnome-desktop camp! ;) But that is no problem for us that doesn't do contract programming and can choose our own tools.

    Maybe they only are afraid to stick out their neck without the warm fuzzy feeling of consent of others?

    I will remember that you too didn't not dare to venture an early opinion!

    Do you know what a chicken is?
  116. Peter is a careful person[ Go to top ]

    Why D is under the radar is a constant amazement to me. Especially in the gnome-desktop camp! ;) But that is no problem for us that doesn't do contract programming and can choose our own tools.Maybe they only are afraid to stick out their neck without the warm fuzzy feeling of consent of others?

    Yes, perhaps they are. It is called 'a professional attitude'. I would suggest that a less-professional attitude is to produce effectively unsupportable code for a client by writing it in a language few will have heard of. Still, I guess it means that you keep the support contract.
  117. resistance is futile[ Go to top ]

    1) .NET is most used in new projects and

    Er no. You need to actually provide evidence for this. If you look at the jobs you will see the exact opposite: .Net is used a lot in legacy support projects, where existing VB6, Visual C++ and COM+ projects need to be continued, but, of course, Microsoft is dropping support for the development tools.
    I predict that .NET will have overtaken Java before the end of this year. By that I mean that the search: c# or .net will yield more jobs than the search java or j2ee at www.it.jobserve.com before 2005-12-31.RegardsRolf Tollerud

    That is an interesting prediction. We shall see. However, it has little or no relevance to the Java job market. This is why I keep showing the rapidly growing use of COBOL vs. VB.NET in France. COBOL has now overtaken VB.NET, and is used in around twice as many projects. However, to say that it has 'overtaken' VB.NET is meaningless, because they are used for different things. .NET is widely used client side (often with J2EE server side). Java is hardly used client side at all.
  118. mathematics is futile[ Go to top ]

    2) .NET and C# is gaining on Java by ca 1% every month
    ....
    I predict that .NET will have overtaken Java before the end of this year.

    Yet more mathematical hilarity.

    So you say that C# is gaining on Java by about 1% a month AND it will have overtaken Java before the end of this year.

    Well, C# is about 50% of Java jobs on the index you use. It is May. There are 7 months left. Add about 1% per month. That gives ... 50 + 7 = 57% by the end of the year.

    So, sorry, you are wrong. It can't be gaining by about 1% a month AND have overtaken Java by the end of this year. It must be gaining by about 7.2% a month to do that, which it isn't.
  119. the situation May 9 2005 was[ Go to top ]

    java or j2ee = 2324
    c# or .net = 1814

    (1814/2324)*100 = 78% With only 1% gain every month it will reach no more that 85% by december.

    I think though, that the gain will speed up. This month it gained 3%! Those kinds of changes go slowly first, then faster and faster.
  120. the situation May 9 2005 was[ Go to top ]

    java or j2ee = 2324c# or .net = 1814(1814/2324)*100 = 78% With only 1% gain every month it will reach no more that 85% by december.I think though, that the gain will speed up. This month it gained 3%! Those kinds of changes go slowly first, then faster and faster.


    Er. You have been presenting C# vs Java figures to us. By that index C# is at 51%.

    You can't suddenly change what you are measuring.

    Even then, taking a UK jobsite which, after all, only shows a few thousand jobs, as representative of the whole IT world is pretty dumb. All you are doing is showing jobs that are present on THAT jobsite!

    Let me show you what I mean:

    The "French Index" (jobs at monster.fr:

    C# OR .NET: 273 jobs
    Java or J2EE: 715 jobs

    This is showing C# at 30%. (it has also been showing VB.NET as static or falling)

    The population of France is greater than that of the UK, so my figure could be statistically better!
  121. dice.com confirm UK results[ Go to top ]

    "Taking a UK jobsite which, after all, only shows a few thousand jobs, as representative of the whole IT world is pretty dumb. All you are doing is showing jobs that are present on THAT jobsite!"

    Well with my limited brain I can do only so much! :) I have tried before to explain why I am using www.it.jobserve.com. monster.com quite right would be more interesting. But unfortunate this rule they have not to show how many hits if it's more than 1000 destroy the purpose!

    However I have found another American job-site, dice.com. There is some problem with it as it converts your searches unasked for. But if you search for j2ee you get 10957 hits, (it has converted your search to j2ee OR java OR ejb)

    If you search for .net you get 8181 hits.

    (8181/10957)*100 = 75 %

    You see? Pretty close to the UK jobsite result (78 %)

    The French Index was interesting but as you know, the greater number the more statistical significance.

    Regards
    Rolf Tollerud
  122. P.S.[ Go to top ]

    I overlooked that using the "advanced search" I can do the same search as at the UK site, so hera are dice.com results:

    java or j2ee = 10839
    c# or .net = 8559

    (8559/10839)*100 = 79 % almost exact the same as at jobserve.com!

    Let us now follow these two sites in the months ahead.
  123. P.S.[ Go to top ]

    I overlooked that using the "advanced search" I can do the same search as at the UK site, so hera are dice.com results:java or j2ee = 10839c# or .net = 8559 (8559/10839)*100 = 79 % almost exact the same as at jobserve.com!Let us now follow these two sites in the months ahead.

    Why? What is this single figure supposed to show? Please read my post about this (just above this one).
  124. "C#/.NET jobs are adding new functionality to, or replacing, existing Windows/COM projects with legacy C/C++ code"

    Please Steve I have been living with this .NET business in three years and I can ensure you that nobody is making windows applications in vb.net c#.net. The little windows development that still are done is in VB6 or unmanaged C+ which does not require any preinstalled libraries and gives much faster applications. The .NET market is the Web - that’s the only thing it’s good for.

    Now we have two large sites, one in US, one in UK that shows the exact same numbers with the same search. The interesting thing to see is the direction the both sites take in the months ahead. If they move in tandem in the same direction even you must admit that it has some significance. And my guess is that the java/J2ee and the C#/.Net job market will be of the same size around the end of the year.

    Regards
    Rolf Tollerud
  125. "C#/.NET jobs are adding new functionality to, or replacing, existing Windows/COM projects with legacy C/C++ code"Please Steve I have been living with this .NET business in three years and I can ensure you that nobody is making windows applications in vb.net c#.net.

    Absolute nonsense. Just go to jobserve and look at the actual jobs: I see on the first page a job wanting C# + UML + C++ + COM, another wanting C# + WinForms for rich client. Another for C# + Front Office. Another is C#/VB, upgrading VB/COM/C++ to C# - exactly as I described.

    Obviously, you are plain wrong.
    The little windows development that still are done is in VB6 or unmanaged C+ which does not require any preinstalled libraries and gives much faster applications. The .NET market is the Web - that’s the only thing it’s good for.

    Again, nonsense. Look at the actual jobs! Don't take my word for it - the evidence is there for you to read. Saying .NET is only good for the Web is very silly.
    Now we have two large sites, one in US, one in UK that shows the exact same numbers with the same search.

    And I showed another (bigger) site with completely different figures.
    The interesting thing to see is the direction the both sites take in the months ahead. If they move in tandem in the same direction even you must admit that it has some significance.

    No, not unless you can give an explanation of what the numbers mean.
  126. do not compond to the folly[ Go to top ]

    "And I showed another (bigger) site with completely different figures."

    You did not. You used a different search.

    "When you make a fool out of yourself, you should at the very least try to keep a certain amount of dignity and aplomb, and not compond to the folly".

    Good advice.
  127. do not compond to the folly[ Go to top ]

    "And I showed another (bigger) site with completely different figures."You did not. You used a different search. "When you make a fool out of yourself, you should at the very least try to keep a certain amount of dignity and aplomb, and not compond to the folly".Good advice.

    Yes I did. I showed the aggregated job site which, with its wildly different C# figure, proved that your analysis was nonsense.
  128. It is all nonsense[ Go to top ]

    Just to show much these things vary, and why you should not take them seriously - indeed.com is a job site aggregator - it searches other jobs sites for you, so has a very large set of information.

    Today:

    Java 59,223
    C# 17,842

    Ratio: 30%

    This differs wildly from the 51% in the UK job site.

    This is why these figures are meaningless.
  129. What on earth do you mean? Why don't you use same search as on jobserve.com and dice.com?

    Results for indeed.com,
    java OR j2ee = 65994
    c# or .net = 60823

    (60823/65994)*100 = 92 %

    Thank you for the tipping me of the indeed.com site.
    Let us include even that one in the coming months!

    Regards
    Rolf Tollerud
  130. What on earth do you mean? Why don't you use same search as on jobserve.com and dice.com?Results for indeed.com,java OR j2ee = 65994c# or .net = 60823 (60823/65994)*100 = 92 % Thank you for the tipping me of the indeed.com site. Let us include even that one in the coming months!RegardsRolf Tollerud

    You are showing that you have absolutely no understanding of this!

    If one site shows C# at 51% of Java, and another shows it at 30% of Java, comparing .NET as a whole is totally meaningless, because the difference in C# figures shows that you aren't comparing the same thing. You know full well that .NET is a large range of technologies, including ASP, and even Java!

    All you are showing is your utter lack of understanding of statistics and any ability to rationally analyse data.
  131. not all sites can search on C#[ Go to top ]

    Because indeed.com is an aggregator we do not even know which sites it search on. In the past there have been sites that didn't take kindly to the # character and refuse to search on it.

    But finally we have a search java or j2ee and c# or NET that works for 3 large sites. Sorry on your behalf!
  132. P.S.[ Go to top ]

    dice.com results:
    java or j2ee = 10839
    c# or .net = 8559

    It's hard to believe that Java has more job openings than all the .NET stuff (including VB, ASP and all client and server side coding). Considering how many VB programmers there used to be, that's simply astounding that Java has more traction.

    I wouldn't believe it if I hadn't seen it with my own eyes ..

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  133. pretty good for a Java fork[ Go to top ]

    "It's hard to believe that Java has more job openings than all the .NET stuff (including VB, ASP and all client and server side coding). Considering how many VB programmers there used to be, that's simply astounding that Java has more traction.

    I wouldn't believe it if I hadn't seen it with my own eyes .."


    Have no fear Cameron, the situation will not last much longer! :)

    And BTW, "nobody’s using it!". Do you remember who said that?

    hi hi (i prononced like i in achilles)

    Best regards
    Rolf Tollerud
  134. From the OOP expert...[ Go to top ]

    And BTW, "nobody’s using it!". Do you remember who said that?

    "AutoCAD is OOP's ancestral beginning"

    Remember who said that? :)
  135. From the OOP expert...[ Go to top ]

    "AutoCAD is OOP's ancestral beginning"

    Remember who said that? :)

    Yes it was me and I was right. ORM is nor OO.
  136. From the OOP expert...[ Go to top ]

    "AutoCAD is OOP's ancestral beginning"Remember who said that? :) Yes it was me and I was right. ORM is nor OO.

    Please don't compound your original mistake! You know full well that

    1) OOP did not start with the AutoCAD drawing program in the 80s.
    2) Neither the AutoCAD software or it's development tools use Object Oriented Programming. AutoCAD development is largely in plain LISP, with no object capabilities.

    So, no, you weren't right.

    ORM is not OO in itself, as ORM is a mapping between Objects and Relational. Hence the 'M' for 'Mapping in ORM'. However, it can be used in an OO way. And, this is wildly off topic and has nothing to do with AutoCAD, VB, Java or anything.
  137. life is hard[ Go to top ]

    Steve, this is difficult for me as I like to encourage and embolden people, not press them down! I understand your wish to be "a real OO programmer" and why should I strangle the budding hope?

    "ORM is not OO in itself, as ORM is a mapping between Objects and Relational. Hence the 'M' for 'Mapping in ORM'

    Yes that is right, and the only thing you ORM people then do is to present the "getters and setters object" on the form, almost finished. Save and New. Finished.

    Why don't you read Alexanders Jerusamlem's clarification once more?

    "For data centric business systems, separation of code and data is a necessity. For a windowing system or a flight simulator or a computer game or a graphics package, encapsulation is a desirable means for managing dependencies. These applications own their data, but business applications don't"
    http://www.theserverside.com/news/thread.tss?thread_id=33576#168623

    "computer game or a graphics package" (Not that a web programmer in a 1000 year could produce a game, even a simple one).

    Ack, now I have destroyed a persons "life lie" again. (Ibsen)

    I never learn.

    Sorry
    Rolf Tollerud
  138. life is hard[ Go to top ]

    Yes that is right, and the only thing you ORM people then do is to present the "getters and setters object" on the form, almost finished. Save and New. Finished.

    Where did you get that silly idea? Very little ORM is used like that. Mostly it is used for mapping relational data to objects for financial calculations or other forms of data processing. For example, I use ORM on one website to extract data for some complex mathematical processing which is then presented as a summary to the user.
    Why don't you read Alexanders Jerusamlem's clarification once more?

    Because I disagree with it. Can't you think for yourself?
    I never learn.

    True.
    Sorry Rolf Tollerud

    I will accept your apology if you stop going round in circles and make some effort to learn.
  139. Ah, the complex mathematical processing, I forgot the complex mathematical processing!

    How could I do that?

    I apologize a 1000 times!
  140. Thanks for the apology[ Go to top ]

    Ah, the complex mathematical processing, I forgot the complex mathematical processing!How could I do that? I apologize a 1000 times!

    The point was to illustrate the obvious fact that ORM is obviously used for far more than just acting as a simple intermediate layer between web interfaces and databases. I assume you may be working on projects where no processing or transformation of the data between the store and the user interface is required. My work is more complicated. I can't give specifics for my actual projects - that would not be professional - but I can give you an analogy. Suppose you are searching for a possible flight an airline. You will need to specify some attributes of the flight - rough price band (what class), number of travellers, times of day etc. The application will need to search for flights that meet your requirements. There may be deals or special offers, and combined with your search requirements some pretty complicated calculations may be required. The application may have to interact with services in other companies to confirm booking possibilities and prices. A lot of state may have to be held on the server as you browse through the site, refining or changing your requirements. This is obviously FAR more than 'getters and setters of ORM objects interacting with forms'. Also, it can't be done with rich clients because there are millions of potential customers with a very wide range of browsers, and there could be potentially thousands of active sessions at any one time, so it makes sense to cache objects in memory, to avoid a phenomenal amount of round trips to the database; your strategy of rich clients and stateless servers would be impossible to use. I work on systems that resemble this. I have, of course, simplified things for you.

    So, you are wrong about how ORM is, and can be, used. I accept your apology.
  141. But I suggest we stop here. I have no wish to humiliate another human being. After all you are worth 100x other worms in this thread no name mentioned! Enough to say that IMO your little problem can be solved more easily with SQL queries and computations.

    But the good thing with these last comments is that we found dice.com and indeed.com! That is indeed very valuable! :) Now we have more meat on our proverbial statistical bone. Definite, particular, exclusive and sufficient proof! That will be our armor against darkness and fud and people that are more interested in hiding truth than finding truth in the months ahead.

    Regards
    Rolf Tollerud
  142. But I suggest we stop here. I have no wish to humiliate another human being. After all you are worth 100x other worms in this thread no name mentioned! Enough to say that IMO your little problem can be solved more easily with SQL queries and computations.

    Then you have not read what I have posted. If I resorted to SQL queries for everything the web services would collapse under the load. These services had already been written this way, and they failed! Re-writing with ORM and processing with objects in memory provided hugely improved performance. Did you not read what I wrote about avoiding round trips to the database?
    But the good thing with these last comments is that we found dice.com and indeed.com! That is indeed very valuable! :) Now we have more meat on our proverbial statistical bone. Definite, particular, exclusive and sufficient proof! That will be our armor against darkness and fud and people that are more interested in hiding truth than finding truth in the months ahead.RegardsRolf Tollerud

    Completely the reverse. We have nothing of any statistical use at all. That was clearly shown when indeed.com showed a dramatically different proportion of C# jobs.from the jobserve.uk site. Do you realise what this means? It shows either a huge sampling error, or it shows that the jobs presented at the sites are not for the same thing at all, and so can't be compared. This is irrefutable proof that the ratios have no useful meaning. I also have provided direct proof (from the contents of job adverts) that a significant number of the .NET/C# jobs are on Windows clients and not competing with Java.

    This means that we can safely ignore any further 'Yann Caroff' type figures you present as having no useful information content. I

    I realised that the apology was intended as sarcastic, but I had hoped that you would have had the good manners to allow me to ignore that.
  143. Do you remember who said that?hi hi (i prononced like i in achilles)Best regardsRolf Tollerud

    There's that man/girl again..damn, wonder what he(?) likes to call her..
  144. did I hear something?[ Go to top ]

    Mikael? I know you are a student right out of school. Does your mom know that you are here? I will answer you but show the permission from your ma first.
  145. did I hear something?[ Go to top ]

    Mikael? I know you are a student right out of school. Does your mom know that you are here? I will answer you but show the permission from your ma first.

    Considering your posts in this forum and other forums, I'm glad I have a somewhat fresh education(5y old). The day I start writing as much gibberish as you do, I'll go back to school instead of making a complete fool of myself. Enough said.
  146. is that really necessary?[ Go to top ]

    Must we use flawed data to draw conclusions? I think it is safe to say the following about job opening data.

    1. any sort of aggregation across sites is not meaningful without filtering out duplicates

    2. going by the job description is not a true reflection of what the platform really uses

    3. any sort of pattern grow/decline should take 5-10 years of data into account. it should also take into account seasonal changes and take that into consideration.

    4. economic and other environment factors must be taken into account in the analysis

    5. old posting that are no longer open should be filtered out of the results

    Now the real question is this, "have analysts gone to this extent to measure the real growth or decline?" None of the reports I've seen the last 10 years have gone to this extent. The reason is it's damn near impossible, since the data is dirty. Drawing any sort of meaningful conclusion from flawed data is humorous at best. A better method of determining actual growth/decline is to do a huge survey of 10K businesses and get a detailed account of the applications, the staff and development.

    enjoy

    peter
  147. but what does it confirm?[ Go to top ]

    "Taking a UK jobsite which, after all, only shows a few thousand jobs, as representative of the whole IT world is pretty dumb. All you are doing is showing jobs that are present on THAT jobsite!"Well with my limited brain I can do only so much! :)

    This is my point. To get a true picture would require a hell of a lot of research and analysis. Just picking one ratio off one jobsite produces a figure which is very hard to interpret. So why are you bothering? Your figures may show something, but it is very hard to figure out what.
    I have tried before to explain why I am using www.it.jobserve.com.

    Yes, because it has fewer jobs. But that makes the results far less meaningful, as you would agree.
    monster.com quite right would be more interesting. But unfortunate this rule they have not to show how many hits if it's more than 1000 destroy the purpose!However I have found another American job-site, dice.com. There is some problem with it as it converts your searches unasked for. But if you search for j2ee you get 10957 hits, (it has converted your search to j2ee OR java OR ejb)If you search for .net you get 8181 hits.(8181/10957)*100 = 75 %You see? Pretty close to the UK jobsite result (78 %)

    But what does this actually mean? Before you were talking about the change in percentage. Now you are comparing absolute percentages. For all we know, the dice.com figure could have been static for months, or even falling! I'm not saying it has been, but who knows?
    The French Index was interesting but as you know, the greater number the more statistical significance.RegardsRolf Tollerud

    Yes, I was sampling jobs from a higher population.

    But seriously - for any statistic you give, I can pick one which shows the opposite trend (e.g. COBOL overtaking VB.NET).

    We have been through this all before. There could well be a huge increase in C# jobs with no impact at all on the Java job market because (as is the case) a lot of those C#/.NET jobs are adding new functionality to, or replacing, existing Windows/COM projects with legacy C/C++ code. .NET is designed for that; Java isn't. Growth in that job area is irrelevant to the Java market. There can be (and I'm sure there is) a large growth in Java jobs that does not impact .NET jobs. For example, the number of Java projects on Linux is huge, but .NET is hardly used for that at all.

    I'm sure that some C# are taking away some Java jobs, and vice versa, but how many? Almost certainly a major part of the growth of .NET has nothing to do with Java at all, but is a simple upgrade of existing VB6/VC++ development. Simple job figures don't help us to understand this, because if you look at the actual adverts you see that there a lot in which Java and .NET are used together! J2EE on the server, .NET for Windows clients.

    So, to continue posting a single ratio is pointless unless you are prepared to put in a LOT of research and analysis to find out what that ratio (and changes in it) actually mean.

    This is why people hire expensive consultants and analysts!
  148. but what does it confirm?[ Go to top ]

    This is my point. To get a true picture would require a hell of a lot of research and analysis.
    Right. The other thing about job postings is that there are many different postings for the SAME job. So one would have to factor those out to get an accurate picture. Also, usually the best Jobs are usually not posted. Also, who knows how out of date the postings are. Some companies are pretty lazy about removing postings.

    So Rolf, the "statistics" (using the term VERY loosly) you are gathering mean little, and even less in light of your argument.
  149. but what does it confirm?[ Go to top ]

    Oh, and a lot of postings actually have both Java and .Net (etc.) in them because they sometimes say having one is ok or having one is good. But the job isn't actually for that language/platform.
  150. dice.com confirm UK results[ Go to top ]

    However I have found another American job-site, dice.com. There is some problem with it as it converts your searches unasked for.
    Because it is a site for finding jobs - not for coming up with useless and meaningless statistics.

    And since most job posters at that site are recruiters and HR people and they usually have little understanding of what they are posting, the job board has done a little interpretation.
  151. The answer is of course, that it is not allowed to change/delete important fields. User defined fields do no belongs in that category but is still very useful.

    Uh, for most comapnies, if you capture data, it's important, but I don't need to explain this to someone as smart as yourself.

    Since Rolf can not take this discussion seriously, I think we better set a point for it right now. At least I put a stop to answering anything that is not at least moderately intelligent!
  152. Michael don't be sad. To be smart is not everything! You can still be an agreeable person, to go out with, have a glass of wine, laugh, get it?

    Am sure the Cedric is pretty smart but nobody wants to go out with him! (Except Cameron of course. Think of Cedric and Cameron being out together! The teacher and the pupil..)

    Jack, now I need the snickering, giggling signs again, should you not help? Since nobody is helping me with that I have to use my own language again! hi hi
  153. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?

    --
    Cedric
  154. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?-- Cedric

    +1
    (well, obviously)
  155. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?-- Cedric
    +1(well, obviously)
    +1
  156. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?-- Cedric
    +1(well, obviously)
    +1

    +200
  157. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?-- Cedric
    +1(well, obviously)
    +1
    +200
    +1
  158. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?-- Cedric
    +1(well, obviously)
    +1
    +200
    +1

    +1
  159. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?-- Cedric
    +1(well, obviously)
    +1
    +200
    +1
    +1
    +1
  160. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?-- Cedric
    +1(well, obviously)
    +1
    +200
    +1
    +1
    +1
    +1 (Duh!)
  161. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?-- Cedric
    +1(well, obviously)
    +1
    +200
    +1
    +1
    +1
    +1 (Duh!)
    +1 Duh +1
  162. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?-- Cedric
    +1(well, obviously)
    +1
    +200
    +1
    +1
    +1
    +1 (Duh!)
    +1 Duh +1
    +1 (salsforce.com runs on Resin)
  163. to kick and scream[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?-- Cedric
    +1(well, obviously)
    +1
    +200
    +1
    +1


    +1
  164. What fun[ Go to top ]

    Haven't enjoyed a series of postings so much in ages. Congrats all round. When they make the film, who will Vin Diesel get to play?
  165. is Cedric a former has been?[ Go to top ]

    As for all these repeated attempts to picture me as an idiot, ninja, unprofessional etc do you really think people buy it?
    I do. Anybody else?

    So I have offended Cedric. A thousand apologizes!

    Strange, Cedric is not the first person I personally would choose to attack. He has on several occasions before demonstrated that he is able to think for himself. For instance when people proclaimed 2004 as "the year of Linux", Cedric pointed out a number of reasons why it would be more appropriate to call it "the year of windows". And in this thread his views about components and "dll hell".

    That said, I have to say that he is (probably) not the kind of guy I would want to socialize with privat. Why not? Well, as the British say it seems that "his hat has become too big". That is pretty obvious both from his comments in this thread and his reaction one time when I happen to misspell his name.

    At Weblogic he did Java programming - I am pretty sure that at Google they will not touch Java with a 10-foot pole even! So what do he do? His bLog is not interesting anymore.
  166. is Cedric a former has been?[ Go to top ]

    I am pretty sure that at Google they will not touch Java with a 10-foot pole even!

    Wrong yet again.

    From a job advert at on yahoo posted this month:

    Google is hiring Java developers...
    "We are hiring senior Java software engineers to write server-side code for web-based applications, for both internal and external use."
  167. Not only that but modifications of the names of Tables and Columns should not affect the application either. And keep as close as possible to SQL ANSI 92 so changing the database cause as little problems as possible. At last, finally and very important: always store your dates in string format, 'YYYY-MM-DD'.

    I can think of a couple of ways of isolating the application from database changes.

    1. use some kind of mapping, so that the code only uses the mapped name. for example, I might have first_name, middle_name, and last_name. If I always use first, middle, last in my applications and have some kind of "CRUD" handle it, it should be ok.

    2. use sql to give the table and columns an alias. similar to the first approach, but instead of some code outside the database, you're using Sql. example: select mytable.col1 name, mytable.col2 last from mytable.

    That's all good for the simple cases, but lets say the end user can define a completely new object and new concepts in the tool. It's going to be rather hard to make it so the application doesn't need to be changed if tables and views are mapped to a specific object.

    It is possible, if the design is primarily map driven so that it doesn't use strong typing. In other words, the application doesn't map a domain object to a class object. That's a valid approach, but even in that case, having something like runtime discovery may be useful. So far, most people come up with their own approach, but I don't see how having a spec to standardize these kinds of operations is bad.

    I've designed and written language transformation frameworks in the past several times and these techinques are quite useful. As others have stated adding a column shouldn't require changes to the application. But deleting a column or deleting a table won't be quite so simple. In many integration application, the core functionality is what you call "CRUD". I can understand how simple applications with 1 database and 1 model consider it CRUD. when there's multiple models and database, the CRUD is what makes or breaks the entire application.

    enjoy

    peter
  168. traumatic experience[ Go to top ]

    That has nothing to do with the ten very important areas where I showed Cameron to be wrong, and mark my words, I was just varming up.

    It's pretty obvious that you were just warming up, since you forgot a couple of little things:

    1) To quote what I said, and ..

    2) To prove that it was wrong.

    I'm sure you've figured out by now that I am just trying to waste your time, since you will be unable to achieve those two very simple steps above.

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  169. traumatic experience[ Go to top ]

    I am making applications, killing client apps Rolf.
    How many apps have you killed? ;)
  170. so on and so on, I could go on forever.
    Sure you could, this list is just an enumeration of exaggeration, fabrications and out of context quotations.

    But it doesn't matter. Whether you like it or not, Cameron is respected here because he has shown his expertise on more than one occasion. I am still trying hard to figure out what *your* technical expertise is, besides being a rabid advocacy poster and expert troll, that is.

    --
    Cedric
  171. the truth is all that matter[ Go to top ]

    what does it matter what one do or not do?

    But Cedric, in this thread just on this occasion you have temporary stepped into my role. What you write about VB components is the plain truth IMO, yet nobody agrees with you and you are attacked. How does it feel? To be right and be persecuted? Just like Dr. Semmelweiss in Vienna in 1847!

    Isn't the truth the all above important?
  172. the truth is all that matter[ Go to top ]

    But Cedric, in this thread just on this occasion you have temporary stepped into my role. What you write about VB components is the plain truth IMO, yet nobody agrees with you and you are attacked. How does it feel? To be right and be persecuted? Just like Dr. Semmelweiss in Vienna in 1847! Isn't the truth the all above important?
    You know, this underlines fundamental differences between you and me.

    First of all, I never think that I detain the absolute truth. I don't even believe there *is* an absolute truth. Only personal experiences.

    Second, I don't feel attacked when somebody disagrees with me. I just think... well, "interesting, this person disagrees with me".

    I never learn something when somebody agrees with me, so I welcome dissent, especially when it's worded respectfully and that the person makes good points (which the posters above have made).

    --
    Cedric
  173. the truth is all that matter[ Go to top ]

    Ah, excuse me, I forgot that in your relationship with Cameron you are the "Pupil".
  174. what does it matter what one do or not do?
    If what one do doesn't matter, why the hell are you wasting your time here? Or do you believe you're achieving anything by pestering this forum like this?
    Isn't the truth the all above important?
    In your case, trolling seems to take precedence over anything else.
  175. I produce faster. I produce smaller.
    Are you still talking about coding? Not sure, because I have heard rumors ...
  176. I know that I any time can write an application that beat the hell out of your. That is valid for all aspects, speed, user functionality and pure beauty. That I can prove anytime. Please take up the challenge. I produce faster. I produce smaller.

    Rolf,
    You act and talk like a 15 year old. I have no problem with you disagreeing with anyone,but don't waste our time. I haven't seen you post anything yet that helps anyone get anything done. You obviously don't have a job, because you spend all your time arguing about obscure, irrelivent crud.
    But, that's all I have time for, I gotta get back to work.
  177. Walter Bright[ Go to top ]

    <blockqoute>I have not commented on Walter Bright. I am sure he is bright, no pun intended, but I've never heard of him before this thread.
    </blockqoute>

    He wrote the lean and mean Datalight C compiler for MS-DOS a few geological ages ago. I bought a copy at the time because I liked to collect such things.

    Walter Bright
    Engineer, Digital Mars
    http://www.cmpevents.com/SDw4/a.asp?option=G&V=3&id=228826
  178. ?[ Go to top ]

    Throw shit at Charles Petzold! Comparing some morons at Sun with one of the worlds greates geniuses!
    Rolf, if you think that Petzold is a genius, then you've pretty much explained the extent of your technical capabilities, as if you hadn't already made your lack of ability abundantly clear.Peace,Cameron PurdyTangosol, Inc.Coherence: Cluster your POJOs!
    I think Rolf is trying compensate it by shooting opinions, quotes and verbal abuse at all directions. Pity us.
  179. Are you serious?
    Yes.
    Creating a component in VB is a breeze. Note that I am talking about the *packaging*, not the API, which is what really defines a component.Write your code in VB, export a DLL or an OCX and the rest of the planet can use it.
    I wish it was that easy. (btw, dlls are not for components). In VB you had to create UserControls. What a pain. If you developed them for yourself, you really needed a separate machine. The same went for dlls. To top it off, try refactoring one of those babies. Just watch your registry fill up.

    Lets see, how do I do that in Java? Ummm. Export as jar. MAN, that was tough.

      No need for them to recompile it, build it, update it or whatever. Just take a look at the IDL interface and you can invoke it from any language.HTML renderer, ftp object, graphic controls, etc...
    there are hundreds of these out there and I bet quite a few on your own machine, which you can discover and reuse with the OLEBrowser.
    How many built with VB? very few. How many built with C++? Most.
    Reusable Swing components?None.
    http://weblogs.java.net/blog/hansmuller/archive/2004/10/and_then_there.html

    Cedric, I have been doing VB development since 94. I have taught VB classes and mentored other developers and teams at multiple companies on VB dev. I got my VB Desktop MCP in one try without studying a lick. :)
  180. If you love Swing framework, but need your app to run with standard internet browsers you could check out http://demo.vaadin.com/sampler/ . You create applications using Java (the Swing way) and because of that all your GUI code can benefit of the full power of Java language such as composition and inheritance, let alone the stuff that good IDE's offer like refactoring UI code. What I am really waiting for is stable version of Vaadin WYSIWYG editor for Eclipse. Remember the good old times when Visual Basic came out?-) Hope this turns out even half that nice of a tool.. Disclaimer: I'm a developer that dislikes declaring GUI stuff using XML or HTML page based "code" + I've done bunch of Swing and Vaadin coding.. Cheers!
  181. Shouldn't that read "Threatens" instead of "Promises"?

    :-)
  182. Shouldn't that read "Threatens" instead of "Promises"?:-)
    LOL!
  183. programming for real men[ Go to top ]

    Sorry guys, the train has gone.

    The future of client programming (as long as there is going to be any demand for clients apps) belongs to the D programming language.

    Check it out at,
    http://www.digitalmars.com/d/index.html

    So is there going to be any demand for client apps? I think so. The BIG question "Rich clients or browser applications" is going to be answered with "BOTH". A first class application in the future will come in two versions, a specialized brower version and a specialized client version.

    Regards
    Rolf Tollerud
  184. programming for real men[ Go to top ]

    Rolf, sometimes it is really hard to understand what you are trying to say.
    What the new language has to do with this topic ?
    Building client side apps is not about language.
    It is about tools.
    And besides, from what i've seen on their site this new D language does not have practically any advantages over well established python and ruby.
    And lacks heavy industrial tools support C# and java have.
  185. first victim will be Mono[ Go to top ]

    Back in old days VB was too easy, but C/C++ was too complicated and cumberstoned. Now we have Java and C#. But these languages come with a VM. Even if you can compile to native it is difficult and you do not really win any speed and even a simple "hello world" app takes up many MB. Also these languages are too easy! (No challenges for us real men).

    Then Walter Bright introduces D, check out his credentials! http://www.walterbright.com/

    Compiles to native as easily as Java/C# compiles to byte code: dmd helloworld.d

    Small files (50-100 Kb with no dependencies!), speed record at The Computer Language Shootout. Interface nicely with win32, gtk and qt. Garbage collected but you can bypass it if you choose to and lots of lots of features. This is nothing made home by a talented student, whatch a professional genius in action!

    In complexity it is between Java and C++. There is a lot to learn but you never feel you waist your time. I am completely won over. It has tremendous potential.

    Regards
    Rolf Tollerud
  186. first victim will be Mono[ Go to top ]

    Back in old days VB was too easy, but C/C++ was too complicated and cumberstoned. Now we have Java and C#. But these languages come with a VM. Even if you can compile to native it is difficult and you do not really win any speed and even a simple "hello world" app takes up many MB.

    Strange that you have never heard of the many native code compilers for Java, as they have been around for years. GCJ is a good one.
    I am completely won over. It has tremendous potential.RegardsRolf Tollerud

    Poor Ruby On Rails. You were so keen on it. But it is now SO 'last month'. I guess fashions and trends change so quickly.
  187. I though we were talking about client apps? Ruby on Rails is a very good web-applications framework.

    Difficult to keep it apart? Be patient,
    it will come to you.

    Tip: With the D programming language you need to have an extensive knowledge of the GUI API of you choice, Win32, gtk or qt.

    Regards
    Rolf Tollerud
  188. I though we were talking about client apps? Ruby on Rails is a very good web-applications framework.

    You were comparing D with Java, which is mainly used server side.
    Tip: With the D programming language you need to have an extensive knowledge of the GUI API of you choice, Win32, gtk or qt.RegardsRolf Tollerud

    Horrible. How old fashioned! You don't need such knowledge with Java, even with native APIs such as SWT.
  189. Swing?[ Go to top ]

    "Horrible. How old fashioned! You don't need such knowledge with Java, even with native APIs such as SWT"

    Do you know what it feels like? Like pasta that has cooked too long.

    Some of us prefer "al dente".
  190. Swing?[ Go to top ]

    "Horrible. How old fashioned! You don't need such knowledge with Java, even with native APIs such as SWT"Do you know what it feels like? Like pasta that has cooked too long.Some of us prefer "al dente".

    Well, if you want to hack Win32, good luck. Some of us have moved on from that stage.
  191. Yes I do I do. Neither VB, C++. Java or C# has made me entirely comfortable the way D does.
    Neither have I ever had the 100% confidence in anyone the way I have in Walter Bright.

    I've found my habitat.
    I am home!
  192. Yes I do I do. Neither VB, C++. Java or C# has made me entirely comfortable the way D does. Neither have I ever had the 100% confidence in anyone the way I have in Walter Bright.I've found my habitat.I am home!

    I hope you are on Walter Bright's life insurance policy. It sounds like you're banking a lot on a single individual.
  193. The language is more or less finished. You do not think that the great players can stand still and do nothing?

    What was wanted was a language that can compile both to native and to byte code. But compile to native is difficult and compile to byte code is easy. Therefore you have to start in the right end.

    Not the other way around.
  194. The language is more or less finished. You do not think that the great players can stand still and do nothing? What was wanted was a language that can compile both to native and to byte code.

    Java has been doing this perfectly well for years. GCJ compiles Java directly from source code to native machine code. The Toba compiler was doing this way back in the late 90s.
    But compile to native is difficult and compile to byte code is easy.

    Nonsense. It is exactly the same. Both are machine codes, it just that one machine code is run on a virtual machine. 'byte code' IS a machine code. The machine code for the VM.

    I have to say that I like the look of D, and it would be great to have another language like that on the JVM. But D is not good for the reasons you specify.
  195. Nonsense. It is exactly the same. Both are machine codes, it just that one machine code is run on a virtual machine. 'byte code' IS a machine code. The machine code for the VM.I have to say that I like the look of D, and it would be great to have another language like that on the JVM. But D is not good for the reasons you specify.
    Agreed. D looks like an incremental work over C, but not much more.

    If you are looking for a real 5GL language, take a look at Fortress (I wrote a quick review here).

    I don't think it will take the world by storm but it's pioneering a lot of novel and interesting ideas that are most likely a good illustration of the things to come in programming languages.

    --
    Cedric
  196. Cedric,
    "Agreed. D looks like an incremental work over C, but not much more"

    Perhaps so, but the devil is in the details.

    There is an endless discourse in history, what is more important in art, originality or quality?
    I have always been a staunch supporter of quality. Therefore Bach is the greatest composer that ever lived IMO.

    Walter has done a practical work in first hand: realizing that all languages that has been popular has had close similarity to the C, he quietly and unobtrusive did the best compiler ever done. Word record!

    Walter is the equivalent of Johan Sebastian Bach in the computer world.

    Perhaps you can guide me to the Fortress newsgroup? Who has written the compiler? What place has it in the benchmarks? How many posts is it in the newsgroup?
    (D=23155 posts 5 May 2005)

    Regards
    Rolf Tollerud
  197. he quietly and unobtrusive did the best compiler ever done.
    I am curious to hear how you actually come to this conclusion...

    --
    Cedric
  198. a work of art[ Go to top ]

    Sorry Cedric, I can not prove that. It is only my pleasant experience working with it.

    All things in the world can not be measured and weighted. When we talk about art you must be a connoisseur.
    Who hasn't seen the little shrug an art lover does when he encounter something he doesn't approve?

    But I can give you a little inkling,

    1) They are the smallest files I ever seen
    2) They are the fastest files I ever seen
    3) They are the friendliest language I ever used to interface with the OS.

    Goto http://www.walterbright.com/ and read about this guys track record. Have you ever seen such a thourough work before?

    Do you have another candidate to the world best compiler writer?

    If so please show me that fellows CV.

    Regards
    Rolf Tollerud
  199. Or, put it in another way:[ Go to top ]

    Do you think Walter is another "Well-meaning Impractical Theoretician with a cause" from Sun?
  200. a work of art[ Go to top ]

    Sorry Cedric, I can not prove that.
    Glad we agree. And interestingly, this also applies to most of what you write.

    Nothing wrong with that, by the way, I just wanted you to realize you are expressing opinions, not facts, and that -- as you sure know -- everybody has one.
    Do you have another candidate to the world best compiler writer?If so please show me that fellows CV.

    If the compiler is beautiful, why do you care about the CV of its author?

    When you like a car, do you ask the resume of every single person who worked on it?

    Your logic really baffles me.

    --
    Cedric
  201. Nothing wrong with that, by the way, I just wanted you to realize you are expressing opinions, not facts, and that -- as you sure know -- everybody has one.

    Only opinions yes - but what you forget is that opinions one day becomes something else, facts that stand out like pillars. To make a correct prediction of the future is one of the few satisfactions in life.

    Unfortunately for Cameron it is impossible for him to erase from TSS database all what he said and done. What we express here is not written on water, but in concrete, standing for all time. I have nothing to be ashamed of, in fact I am proud of my old opinions here in TSS from the 3 years I have been here.

    Your statement though, that (Fortran) Fortress from Sun (!) should ever become mainstream will make you look like an idiot some time, mark my words. You saw it here first!

    I can not understand how you can say such a thing, after all I know that you are an intelligent being, one of the best. The only thing comparable I off hand can think of is the Nobel Prize discussion we have here in Stockholm every year. When the Swedish television gathers together some Nobel prize winners for an informal discussion.

    What makes me so astonished is that as soon the good scientists begin to drive over the boundaries of their fields their opinions degrade dramatically - even to the ridiculous.

    It is mystic, isn't it?

    Regards
    Rolf Tollerud
  202. fact-checking[ Go to top ]

    Your statement though, that (Fortran) Fortress from Sun (!) should ever become mainstream will make you look like an idiot some time, mark my words.

    Now, I'm not going to guarantee that Cedric won't ever look like an idiot. But I'm pretty sure that when he said "I don't think [Fortress] will take the world by storm" he kinda was saying that he *does not* think that Fortress will become mainstream.

    -Patrick

    --
    Patrick Linskey
    http://solarmetric.com
  203. Nothing wrong with that, by the way, I just wanted you to realize you are expressing opinions, not facts, and that -- as you sure know -- everybody has one.
    Only opinions yes - but what you forget is that opinions one day becomes something else, facts that stand out like pillars. To make a correct prediction of the future is one of the few satisfactions in life.Unfortunately for Cameron it is impossible for him to erase from TSS database all what he said and done. What we express here is not written on water, but in concrete, standing for all time. I have nothing to be ashamed of, in fact I am proud of my old opinions here in TSS from the 3 years I have been here. Your statement though, that (Fortran) Fortress from Sun (!) should ever become mainstream will make you look like an idiot some time, mark my words. You saw it here first! I can not understand how you can say such a thing, after all I know that you are an intelligent being, one of the best. The only thing comparable I off hand can think of is the Nobel Prize discussion we have here in Stockholm every year. When the Swedish television gathers together some Nobel prize winners for an informal discussion.What makes me so astonished is that as soon the good scientists begin to drive over the boundaries of their fields their opinions degrade dramatically - even to the ridiculous.It is mystic, isn't it?
    RegardsRolf Tollerud

    I haven't looked at D, but from all your posts, you seem to think you know about compilers. So my question to you is this.

    Have you ever written a compiler in your 20 years of programming?

    I'm no experts, but writing a full blown compiler isn't easy. Writing a great compiler is even harder. Back in college I was lucky enough to meet an individual who wrote a commercial compiler with his father before he turned 16. This individual was a computer science major and coding for him was trivial. In other words he could program professionally in a half dozen languages before he started college.

    Having written a few rule compilers, your comments and remarks give me the impression you have zero practical knowledge with writing compilers.

    I'm sure that impression is wrong, so feel free to demonstrate your skill :)

    peter
  204. What makes me so astonished is that as soon the good scientists begin to drive over the boundaries of their fields their opinions degrade dramatically - even to the ridiculous.
    It's happened to you quite a few times as well.
    our statement though, that (Fortran) Fortress from Sun (!) should ever become mainstream will make you look like an idiot some time, mark my words.
    You really need to work on reading what people write more carefully. Here is what I wrote:
    I don't think it will take the world by storm

    --
    Cedric
  205. It's happened to you quite a few times as well.

    I disagree, Rolf has always been focused squarely on his field of expertese.

    I am surprised nobody collected a library of Rolfisms and made T-shirts, coffee mugs, and caps and sold them on ThinkGeek.
  206. Are you sleeping? How can you casually dismiss such a great work? Not even Bjarne Stroustrup has done something similar. Bright is the better man of the two IMO. We are talking of the first (number one) place in The Computer Language Shootout! If we make a benchmark your 10 MB compiled Java program will not even have started when the D program is finished. D have GUI programs of 200 Kb with no dependencies! No MFC DLL!

    Regards
    Rolf Tollerud
    (a href="Converting" rel="nofollow">http://www.digitalmars.com/d/index.html">Converting C++ to D</a> )
  207. Walter Bright[ Go to top ]

    I knew his name sounded familiar. He created the game 'Empire'. Boy did I spend a bunch 'O' time playing that.

    -Pete
  208. first victim will be Mono[ Go to top ]

    I am completely won over. It has tremendous potential.RegardsRolf Tollerud

    Does that mean you will be leaving us soon?

    -Pete
  209. D is irrelevant and off-topic[ Go to top ]

    Java has been on the train for years, D has a lot of catching up to do to get anywhere close e.g. the D language features may appear nice if you like the C/C++ baggage (I don't), however D needs loads of libraries, better security, killer apps and recognition to be taken seriously against Java i.e. who cares.
    Sorry guys, the train has gone. The future of client programming (as long as there is going to be any demand for clients apps) belongs to the D programming language.Check it out at,http://www.digitalmars.com/d/index.htmlSo is there going to be any demand for client apps? I think so. The BIG question "Rich clients or browser applications" is going to be answered with "BOTH". A first class application in the future will come in two versions, a specialized brower version and a specialized client version. RegardsRolf Tollerud
  210. June 7 2005
    Today indeed.com for the first time showed more .net than Java jobs!

    Results for indeed.com last month,
    http://www.theserverside.com/news/thread.tss?thread_id=33754#170421
    java OR j2ee = 65994
    c# or .net = 60823
    (60823/65994)*100 = 92%

    this month (June 7 2005),
    java OR j2ee = 62176
    c# or .net = 63703
    (63703/62176)*100 = 101%

    Yann Caroff March 22, 2002
    To me, .NET is a serious danger that should not be overlooked or laughed at.
    http://www.theserverside.com/news/thread.tss?thread_id=12595#44201

    Regards
    Rolf Tollerud
    ("nobody is so stupid as the group")
  211. June 7 2005 Today indeed.com for the first time showed more .net than Java jobs!

    I said I would not bother to reply to you any more, but I found something amusing: A considerable number of those jobs require both .NET and Java skills... here is an example:

    "We are looking for a Java developer to convert an ASP.NET user interface layer to Java Swing."

    That would come up on your .NET search, but probably not in the way you intended :)

    Statistics without interpretation is meaningless, which, of course, you know. You are simply trolling. However, even I was surprised by an ASP to Swing conversion!
  212. However, even I was surprised by an ASP to Swing conversion!
    Having done both - I'm not. Glad to see someone has come to their senses. :)
  213. It was you who found the second site, indeed.com site remeber? If you follow the first link I wrote:

    "Let us now follow these two sites in the months ahead"

    Results for jobserve.uk site last month,
    http://www.theserverside.com/common/printthread.tss?thread_id=33754#170405
    java or j2ee = 2324
    c# or .net = 1814
    (1814/2324)*100 = 78%

    Results for jobserve.uk site (June 7 2005),
    java OR j2ee = 2268
    c# or .net = 1807
    (1807/2268)*100 = 80 %

    So jobserve.uk continues this trend since more than two years back when .NET was just a tiny percentage of Java. So this shift means nothing? When will it stop then, according to you?

    Regards
    Rolf Tollerud
  214. It was you who found the second site, indeed.com site remeber?

    Yes, because it proved that that you can't get anything meaningful or consistent from these job figures, as this job site showed a wildly different ratio of C# jobs to others.

    So why you keep posting these figures, even though you know this, is beyond me.
    So this shift means nothing? When will it stop then, according to you?RegardsRolf Tollerud

    Of course it means something. It means that the number of Job ads that include the words 'C#' and '.NET' is growing.

    To find out when it will stop (according to me), do the following. Look through the details of all job ads to find those which involve development for Windows using Microsoft tools (good keywords to search for initially are Visual Basic, C++ and ASP). The growth in .NET will stop when about 80-85% of those jobs are converted to .NET. (A few legacy jobs always remain). You may want to add a few percentage of growth each year for general IT industry expansion. This will take a while, but I predict we should be somewhere near that stage by the beginning of the next decade. Meanwhile, Java jobs will grow significantly was well, both for new projects and as legacy C/C++ and COBOL projects are migrated to Java and J2EE, and as the mobile Java games market continues to expand.
  215. This has to be one of the most humorous entries ever. A minor correction in the statistics, France does have a slightly larger population, (just over 200k), but in the 15-64 age groups, i.e where most workers are, the UK has 600K more. But this is nitpicking by a biased Brit ;-)
  216. VB components (prior to VB.NET) are COM components. As such they get registered in the system registry of a Windows computer and are global to the machine.

    VB has some hack mechanisms that are intended to try to manage COM versioning. They don't work very well (as in don't solve all the problem 100%) and a lot of VB programmers never bother to try to understand that COM registration stuff and how/why one would try to manage component versioning.

    I worked for a company that attempted to build and deploy customized versions of a VB application on a Citrix Metaframe cluster. They got into a huge instability quagmire because of this issue. The mechanisms introduced in W2K and XP aren't sufficient to solve this problem. Porting this classic-style VB code to VB.NET and using private assemblies was totally not a feasible option. I rescued them by engineering the ability for VB5/VB6 applications to run in COM isolation from one another by hooking in and intercepting the COM APIs of the VB runtime DLL - hence my strict isolation solution worked with prior existing compiled VB code.

    I'll take the Java CLASSPATH and Java classes any day over VB components. The Java CLASSPATH in comparison to the VB component COM/DLL hell problem makes the Java designers all look like geniuses. (Which is perhaps an ironic assessment giving the grumbling of Java programmers regarding the Java CLASSPATH/class-loader - just goes to show that many of them have evidently never seen just how hopelessly mired Microsoft invented technology can cause things to become.)
  217. DLL hell is not Windows-specific[ Go to top ]

    I rescued them by engineering the ability for VB5/VB6 applications to run in COM isolation from one another by hooking in and intercepting the COM APIs of the VB runtime DLL - hence my strict isolation solution worked with prior existing compiled VB code.I'll take the Java CLASSPATH and Java classes any day over VB components.
    How does the CLASSPATH (a list of absolute path names... seriously) address the versioning problem, exactly?

    It doesn't.

    You have the very same problems in Java that you have with COM and that you have with shared libraries in general. There is nothing in Java that lets you guarantee you are using the right version of the library and there's even nothing to allow you to look up any library at all, for that matter.

    There are also no guidelines and no enforcement coming from the VM to help you solve this problem.

    In these two regards, COM is ahead of anything we have in Java.

    The reason why "DLL hell" is more known on Windows than in Java is because on Windows, end users are exposed to it every day. In the Java world, only Java developers are ever confronted to it and they can usually find their way around.

    The bottom line is: there is no easy way around "DLL hell". It's not a Windows-specific problem and between never using shared libraries or always using shared libraries and appling certain rules, there is an whole spectrum of solutions that all have their pros and cons.


    --
    Cedric
  218. DLL hell is not Windows-specific[ Go to top ]

    Cedric, Have you ever developed and deployed a COM app? One that used common libraries?

    I have. And it is nuts. You install an app and get all the dlls and ocx's right. Then install another app on hope it doesn't break yours because they might require an incompatible dll or just install an incompatible one. And have fun trying to figure out what broke your app. No stack track. No good error.

    The previous poster should have said classpath and NOT CLASSPATH. I do not depend on the System classpath. Because in that instance you are right about dll hell.
  219. DLL hell is not Windows-specific[ Go to top ]

    I rescued them by engineering the ability for VB5/VB6 applications to run in COM isolation from one another by hooking in and intercepting the COM APIs of the VB runtime DLL - hence my strict isolation solution worked with prior existing compiled VB code.I'll take the Java CLASSPATH and Java classes any day over VB components.
    How does the CLASSPATH (a list of absolute path names... seriously) address the versioning problem, exactly?It doesn't.You have the very same problems in Java that you have with COM and that you have with shared libraries in general. -- Cedric

    It does quite simply - the solution to versioning is to avoid sharing modules that are subject to being present in various generations and instead use private copies - as in .NET's private assemblies approach. The classpath makes it very straightforward as a mechanism that can be used to construe a set of Java libraries that are of known origin and version. IOW, Java has eschewed Microsoft's COM concept of registering components in a global registration system for the purpose of sharing them.

    Sure, one could chose to share some group of Java libraries but at any time a private copy - thanks to the classpath approach - can be made use of such that they are of well known origin. Because there's no global registration involved one doesn't have to work out any scheme for requesting a particular version or negotiate hoping to find a version that might be acceptable. All that hokey and senseless complexity is avoided.

    COM compounds its problems because meta information (the typelib) for binding purposes is also registered (globally). This leads to problems when the attempt is made to run different versions of a VB component on the same machine. All the VB developers I've ever been around have not been enough on the ball to know how to try and manage this properly. When one starts dealing in hundreds VB COM DLLs, then it becomes too much even for a those that attempt to do versioning carefully using the native capabilities of COM and VB. Running each VB app in complete isolation, in contrast, is easy and bullet proof - if the capability to pull this off exist (from Microsoft, of course, it does not exist except when choosing to migrate to VB.NET).

    Stability and robustness counts for vastly more than the ability to share code. It took Microsoft a decade to eventually concede that point as with W2K they began to introduce modest isolation features - and then fully conceded the point by providing for rigorous private assemblies for .NET.
  220. finally[ Go to top ]

    It was about time, the design time specification is also a big leak in the JSF specs, basically currently every vendor does his own version, which prevents that you just can plug in something like myfaces into Studio Creator and have a visual component layout.
    In Swing this kind of stuff has ben overdue at least for 6-7 years.
  221. Don't bother trying to win VB developers over there are enough script kiddies doing Java allready. We'd be better of trying to figure out how to get rid of those instead of trying to bring in more!
  222. Don't bother trying to win VB developers over there are enough script kiddies doing Java allready. We'd be better of trying to figure out how to get rid of those instead of trying to bring in more!

    I disagree. People used VB because it enabled them to quickly do things that were valuable for their businesses. I would like to see Java have more capabilities to help people quickly do things that are valuable for their businesses. However, we shouldn't fool ourselves and call all these things "Java" because if these people have to drop down the programming level that Java provides by default, they'll tend to run away screaming. In other words, we need higher level constructs built on the Java platform that provide simple programming options, much like VBA allowed people to customize and script applications. (How to keep it from becoming a mess is a different question.)

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Cluster your POJOs!
  223. Delphi, Workshop and new IDEs[ Go to top ]

    Don't bother trying to win VB developers over there are enough script kiddies doing Java allready. We'd be better of trying to figure out how to get rid of those instead of trying to bring in more!
    I disagree. People used VB because it enabled them to quickly do things that were valuable for their businesses. I would like to see Java have more capabilities to help people quickly do things that are valuable for their businesses. However, we shouldn't fool ourselves and call all these things "Java" because if these people have to drop down the programming level that Java provides by default, they'll tend to run away screaming. In other words, we need higher level constructs built on the Java platform that provide simple programming options, much like VBA allowed people to customize and script applications. (How to keep it from becoming a mess is a different question.)Peace,Cameron PurdyTangosol, Inc.Coherence: Cluster your POJOs!

    Hi,

      Delphi was a class act and did all(most) the right things to fix issues with VB and it came from a respected company(Borland). But Delphi was (only) an implementation which had some specification -- somewhat like Hibernate.

    Delphi had the power that geeks could use PLUS the component architecture and the library that enterprises covet. Even 3rd party vendors started writing Delphi components (including for their Linux port -- Kylix)

      The JCP community(and not just sun) should try to do what Delphi did(and maybe more). Now that Java as a language is mature, this has even fewer risk -- it's just a few JSRs and a new IDE that can build RCP and internet/intranet applications.

      BEA's Weblogic Workshop was an attempt towards that end and if the whole JCP community gets behind this, we can achieve more.

      I hope in a few years, Java will have 1000s of components (yes, 1000s since the range of problems that Java adresses can easily fill hundreds of thousands!!).

    BR,
    ~A
  224. Don't bother trying to win VB developers ...
    I disagree. People used VB because it enabled them to quickly do things that were valuable for their businesses. I would like to see Java have more capabilities to help people quickly do things that are valuable for their businesses. However, we shouldn't fool ourselves and call all these things "Java" because if these people have to drop down the programming level that Java provides by default, they'll tend to run away screaming. In other words, we need higher level constructs built on the Java platform that provide simple programming options, much like VBA allowed people to customize and script applications. (How to keep it from becoming a mess is a different question.)

    I have no problem with less technically inclined people developing for the Java platform as long as they do not pass themself of as Java developers.

    I DO have problems with hacks with a poor understanding of object oriented principles undermining what it means to be a Java developer.

    Too often I am faced with developers who manage to successfully justify their technical incompetence by claiming to be more business oriented.

    I have supported introduction of scripting languages on the Java platform such as Groovy in the hope that it will attract developers who shouldn't be developing in Java in the first place.

    Regards
    Jan Hornbøll Hansen
  225. Back onto Components[ Go to top ]

    OK, if we can forget about clients for the moment and focus on web apps, would it be fair to say JSF will incite [and possibly already has] a move to a component oriented approach to app dev within the java community? I mean, once one switches from a Struts (or the like) mind set to the component oriented JSF, and actually develop a few apps with it (and especially if you've worked with Netbeans 5.5 with Visual Web Pack), I donno, it could be just me, but I kinda think you get the best of both worlds: highly functional and easy-to-use visual components AND java's ability to elegantly structure complexity. Once one transcends the JSF learning curve, at least I think, the productivity AND customisability become obvious. Am I alone?
  226. That whatever comes from Microsoft is A PIECE OF CRAP... not a single frickin product that i know of from Microsoft doesn't suck. Their browser their IM their OS.... the list goes on and on... so while Microsoft is a great company for new innovative ideas.. it will be ultimately driven by bottomlines... so don't mislead people in trying to market their product and build up the hype just that they usually do... open source is the reason why our souls are at peace that we can FINALLY build some stable robust applications.
  227. Re[ Go to top ]

    That is what my title should have been. :) Andy