The future of IBM Visual Age (and other Java IDE's)

Discussions

News: The future of IBM Visual Age (and other Java IDE's)

  1. A new article from IBM implies that they are dropping Visual Age for Java and re-writing their tools to the WebSphere tools line because VisualAge is too "language-centric".

    From:

    http://www7.software.ibm.com/vad.nsf/data/document2020

    "Why is IBM coming out with new tooling?

     ... Now, as the Web application development market continues to evolve, we feel that a new tooling platform is needed. We, as well as our customers, realize the limitations in our current technology and in most traditional, language-centered IDEs. They were not designed for this new environment, and have been adapted and extended to the limits of their capabilities.

    To continue to provide a world-class development environment, we must go beyond an improved Java IDE and develop a new technology base for Web application tools ..."

    Interesting justification in the writeup that language-centric tools are not appropriate. Seems to me that tighter integration with modeling solutions is what would solve this rather than re-writing the IDE.

    Comments on IBM's plans/reasons?

    Threaded Messages (19)

  2. Compete against Microsoft .NET, and offer a similar web service integrated platform & tools suite (like Visual Studio .NET)



  3. Sounds like IBM has has been working with .NET beta and realized MS was on to something. Also, I wouldn't be surpised if some of the Java vendors started working with .NET. I know this is not what IBM is doing in this case.
  4. I think in this context the mentioning of "language-centricism" should not be interpreted as questioning the dominat role of Java in IBMs strategy. It looks to me that they are among the strongest proponents, still, and will remain.

    Rather, I read "language-centric" as "coding-centric", which they seem to want to alleviate with richer tools. And also, they seem to want to do away once and for all with the embedded proprietary JVM in Visual Age. All that sounds ok for me.

    I dont read anything about "dropping Visual Age" - it's being revamped, which makes sense after all this time.
  5. This kind of went off on a different tangent than expected; I had hoped to hear what people think the next generation of Java IDE's will look like (or if it exists).

    I think that is what IBM is suggesting they would like to build (regardless of how good or bad VisualAge is today) and probably what every vendor thinks it already has ...

    For example, a quick survey of the market shows me that:

    * WebGain is trying the "we bundle everything you'll ever need" approach (Visual Cafe,Toplink, Structure Builder, Application Composer, Business Designer ...). Their push seems to be towards 4GL, traditional developers away from the pure Java head ...

    * Borland seems to be "we do the latest J2EE standards first and best" approach; focussing on rich coding as is their heritage and leveraging partners heavily (Rational, CA, Microsoft, Bea, iPlanet ...). Their push seems to be generic Java development but be really good at it and let others plug in.

    * Sun ... open source with NetBeans, and then the Forte enterprise line. Presumably by getting the buy in of the open source community they ultimately get some great extensions and developer support. I assume their push is for the Sun/iPlanet platform but also Java evangelism

    * Oracle is going down a combination of the Borland/WebGain road but instead of purchasing tools companies like WebGain, it is building it all itself, adding built-in UML modeling and code profiling in their upcoming release.

    * Others? Silverstream? Bowstreet?

    Everyone seems to have a start at a web services story with more "coming soon".

    It seems like no one, including IBM, has heavily committed their IDE to Web Services beyond lip service (SOAP proxies, WSDL generation, UDDI browsing ...) as add on utilities ... and this seems like the pretty powerful combo: an integrated J2EE Web Services IDE.
  6. I wonder how much re-education and re-working will be needed with this new IDE. That's a little scary. If IBM would develop a whole new environment because Visual Age is too "coding-centric" it doesn't make sense. However, I do think many developers would rather have a tool with integrated modeling and Java development.
  7. We've been expecting IBM to re-write VisualAge for some time.. the Smalltalk-based codebase is too unwieldy to keep up with JDK changes (perhaps due to staffing at OTI and IBM Toronto labs, not Smalltalk.)

    What I didn't expect was that they'd move away from the interface VA users have grown to love. I really hope this new "WebSphere Application Developer" doesn't turn into a tree-based explorer IDE with a few of VA's feature's rolled in.

    This also doesn't provide a ton of incentive for non-websphere customers to use VAJ. Though this has been an issue for a while with the Websphere Test Environment. IBM needs to take a hint from Borland: gently push your app server offering through your IDE.. don't cram it down our throats.

    Anyway.. here's to what was once the king-pin IDE, now on the altar of sacrifice to the Web Services god.
  8. interface VA users have grown to love


    Hmm... the VA users I have met, myself included, will be dancing about on it's grave singing Hallelujah!

    In fact, this is fabulous news - since we can now say that VAJ is probably doomed in the longrun, we can start opening chequebooks and run, not walk, to a vendor of a decent development environment.

    Cheers IBM, this is the best news I've heard all month.
  9. I keep on seeing reference to models to Java .. what about Ration Rose? As a round-trip engineering tool isn't this suitable for modeling class diagrams and then generating infrastructure code?
  10. From my own personal experience, Rational Rose's two code generation is poor compared to together's control center and others. JMO
  11. You owe it to yourself to try TogetherJ. If you've used Rose then you'll find TogetherJ outstanding++.
  12. "Hmm... the VA users I have met, myself included, will be dancing about on it's grave singing Hallelujah!"

    Based on the results of the last SYS-CON Java IDE survey, where VA was once again voted most innovative product and was the #2 IDE, I don't think everyone agrees with you.

    While I recognize IDE interfaces are a matter of taste, are you seriously suggesting that the typical File-browser/tree-based/text-editor IDE is a superior alternative to VAJ's model?

    VisualAge has one of the most advanced interfaces for code searching, source control, and dependency management. Coupled with jFactor it is a productivity dream. Certainly it's a flawed IDE, but its strengths shine brighter than them.

    But, given the dripping sarcasm of your note, I guess we might as well be from different planets.

    Anyway... I really hope the new WSAD maintains VisualAge's spirit with the less language-centric environment. It still might.
  13. I sense IBM wants to open their environment to accomodate better tools and offer more flexibility to developers. this will improve the current experience developers have today.
  14. "Based on the results of the last SYS-CON Java IDE survey, where VA was once again voted most innovative product and was the #2 IDE, I don't think everyone agrees with you. "

    Being innovative is not the same as being liked.
    I admit their interface is innovative, but I don't like it. Maybe I'm just oldfashioned...
    It is the #2 IDE in sales, but mostly because in large corporations purchase decisions are made at a level where the people who have to use the product have no influence whatsoever. These people get a nice package deal from IBM (things like "buy 10 RS-6000s, get 100 licenses VAJ for free") and go for it. The people downstream then have to use that product whether it is the best tool for the job or not.

    "While I recognize IDE interfaces are a matter of taste, are you seriously suggesting that the typical File-browser/tree-based/text-editor IDE is a superior alternative to VAJ's model? "

    It sure is superior for me. That code-repository and hiding of files in favour of methodbodies just does not do it for me.
  15. My experience is that those who get to know Visual Age really do feel that it is superior to a file-based IDE. The power that is provided by the built-in Envy engine/repository is truly incredible. Does any other product provide comparable ability to make changes to a running application?

    The problem with VAJ is twofold:
    First, it really does take a long time to learn to use VAJ well -- until then it can be very frustrating.
    Second, the comment about VAJ not supporting non-Java source code is right on target -- that is a huge drawback to the current VAJ product.

    The mumbo-jumbo about VAJ sales going to big corporations who don't really care about quality or productivity is absurd. Big corporations do buy in bulk and they do require compliance to a policy, but they invest a great deal of time and expense in evaluating their tools before they do so. Of course, they also place a high premium on stability of the vendor and support too.
  16. I agree about the VA UI...yuck.

    The only thing I like better than tree based class browsers are text editors. Why? Because they don't try to tell me how to code.

    I write smaller, tighter, more maintainable code with a text editor because I *have* to. All these toys just make you fat and lazy. Worse, they make it possible for half-trained monkeys to get into the loop.

    I have JBuilder4 Pro but I seldom use it. I don't like the effect it has on my discipline. I prefer the bar kept good and high. (Real Programmers code in black ink...)
  17. {{{ VisualAge has one of the most advanced interfaces for code searching, source control, and dependency management. Coupled with jFactor it is a productivity dream. Certainly it's a flawed IDE, but its strengths shine brighter than them }}}

    Intellij Idea has all this but doesn't tie you into it's way of working as much as VAJ. VisualAge certainly has some nice features but has some rather irritating issues that spoil the who things. And the speed gets terribly slow when dealing with large amounts of code.

    If you haven't seen Idea then check it out. Version 2.0 was release recently and they've made huge progress since version 1.0.

  18. For me the beauty of VAJ is it's integration with Websphere. Even though it took a while for me to get my head around the 'VAJ-way', I have learnt to love its rich set of features. I would hate to have to go back to having to code with another IDE now, especially for writing j2ee stuff!

    Anyway, at a recent local IBM/Websphere roadshow (Websphere 2001) the IBM guys were adamant that the new 4.0 development tool (Websphere Studio?) would have the option of having a 'VAJ flavour' or a 'skin' that would make it look and smell like VAJ to seasoned VAJ users.

    Also that one of the reasons for the change was to better integrate the EJB development environment into the tool ... despite its great integration with websphere, at the moment it is a little stuck-on and doesn't love the workgroup programming of VAJ that well.

    regs
    scot.
  19. The applications I'm working on are written in Java, HTML, WML, XML, JavaScript, etc.

    JBuilder, Visual Cafe and Forte allow me to work with all the files and languages that make up my app. They may not do color coding and other tricks for each language, but at least I can organize the code consistently, edit it in one IDE without switching out to TextPad, and keep it all in the same source code control system (this is the worst part - how am I supposed to version my HTML/JSP in VAJ? Use two SCC programs?).

    So I see some reasons why IBM wants a less language centric IDE.

    Tom

  20. IBM know they are in trouble. I meet someone recently at a conference who is pretty senior at IBM, and he told me Visual .Net really has shaken IBM. This is their first response.