IBM completed the acquisition of Rational Software in March 2003 and hasn't looked back. In a recent interview, IBM Rational marketing exec Eric Schurr talks about Rational as IBM's "build" platform for its "on demand" strategy, discusses new products, improvements to Rational Unified Process (RUP), etc.
- Posted by: Floyd Marinescu
- Posted on: May 30 2003 16:19 EDT
Read Rational, WebSphere, and You.
- What does having Rational as a part of IBM mean to WebSphere? by Prasoon Choudhary on May 31 2003 04:10 EDT
- What does having Rational as a part of IBM mean to WebSphere? by Andrew Clifford on June 02 2003 11:35 EDT
- What does having Rational as a part of IBM mean to WebSphere? by Wille Faler on June 03 2003 02:56 EDT
The next generation of Java developers will primarily be people who aren't J2EE experts. For those people to be successful, they need technology to help them build applications on J2EE without having to deal with the intricacies of J2EE.
Which generation are you talking about Eric, x, y or z . By the way thanks for entertaining us. Good joke.
Maybe he means the next generation as in the people who program in VB and other 4GL type languages who actualy find it hard to narrow a data type after looking it up and then invoking a method on it. J2EE is hard but its fun so the next generation of tools will ease the development of user interfaces and auto generation of EJBs and mappings to data models. Why not take XFORMS as an example this will invoke webservices and completly removes the need for JSPs and Servlets for certain types of applications. This doesnt devalue what you can do with J2EE it just opens the runtimes to a different market of developers.
I dont think that devaluation is the biggest hesitation, I think its quality.
I have seen some very low quality code from developers who have been guilty of IDE over-reliance for code-generation.
However, hopefully technology has come a long way since the days of VisualAge, Forte 3 etc..
My experience is that it all to often is a case of not just code-generation, its hardcode-generation which is tighly coupled with the target-platform, implementation AND development-tools which is a really bad thing.
I hope the "auto-mapping of EJB:s to data-models" isnt "to (DB2) datamodels.".
Dont get me wrong: DB2 is lovely in my opinion, but its not always that people a) have a choice
b) choose that
If these tools can really generate quality code that isnt coupled to first and foremost the tools, and second intended platform, I think they might have a future.
However, I will be somewhat skeptic but give it the benefit of a doubt until proven in some direction. :)
You might want to spend some time reading the article before you claim a quote which is inaccurate. The quote you have was part of the article that was a description and was not stated by Eric Schurr. That was written by the magazine doing the interview.
Martin Fowler has some interesting insight into UML trends. See http://martinfowler.com/bliki/UmlMode.html. It corresponds with Rational's tool decisions and IBM's decision to buy Rational.
Rational made a decision to tightly couple with two popular IDEs: WSAD and Visual Studio. On the upside, seamless integration to approved tools/services is excellent and enhances productivity. When you want to deviate, those IDEs limit a development teams choice of frameworks, version control, build systems, embedded app servers, test environments, etc..
Trying to use XDE outside of the Websphere mindset is getting tougher. IBM, I assume, will not make it easier. I expect XDE as a plugin to Eclipse will always lag the WSAD environment somehow. Eclipse has everything I want now. Together on Eclipse may be a better bet.
I do find value in protoyping a vertical slice of a system in XDE to communicate a design. Once you know the components to construct in a framework, the UML view has less power.
Thats pretty much what I am worried about..
I have seen some horrible code generated by stone-age ide:s, and to make the least sense of the code (refactor it sensibly), I had to go back to using an IDE that hadnt been supported by its vendor for over 3 years..
And this kind of situation can even backfire on the "lock-in" creating company when their own developers have used these tools.
I am not to enlightened about this particular initiative, so I wont criticise something I no nothing about. However my past experience with rational tools is that they could use a bit of simplification and clean-up, they have been quite horrible to use.
I believe openness to app server, database architecture, development tools, revision control and buildprocess is something of the utmost important. Hey, there are people out there (I´m not one of them) that feel more productive using emacs than some modern IDE, if that is so, why shouldnt we let them?