This week's sold-out
Eclipse Summit
Europe, in Esslingen, Germany, will include
a
presentation on the Topcased project, an effort driven by the
French National Center for Aeronautical
and Space Technology Research to construct "
an open-source CASE environment
[to] perpetuate the methods and tools for critical embedded system
development."
The goals of Topcased include "high-quality open-source model editors
supporting an integrated development process from system specification
to product architecture and implementation," which sounds like a pretty
good idea—but some eyebrows may rise when a reader encounters the
additional goal, "Rely on the long-term structure of the academic world
to ensure project continuity."
I'm not saying that every contact between academia and industry is
necessarily a collision, but these two communities are certainly known
for having differing definitions of success.
When I see a software development tool being driven by higher-level
goals—that is, by someone's idea of how software should be written—I'm at least
as nervous as I am intrigued. In the very early days of the IBM PC, a
Pascal development environment called Alice got a lot of
advertising, with an iconic representation of a tall,
intelligent-looking blonde who resembled an
illustration from a Disneyfied Lewis Carroll story dressed up in a
pinstripe suit. As I recall, though, the dominant theme of reviews of
the product was that a developer who worked with Alice wasn't really
learning Pascal (which was notionally the target language of the
product)—but was rather learning the least cumbersome way to make
Alice generate the Pascal that seemed to do what the developer wanted.
No one seemed to think that this was ideal.
I had the same mixed reaction to last
month's announcement of the release of NetBeans/BlueJ, a version of
the NetBeans open-source Java environment that incorporates elements of a
university-driven research project aimed at teaching object-oriented
development. The Java-oriented
BlueJ is derived in turn from Blue, an
academic language that is as purely object-oriented as a language can
get.
Personally, I'm much more interested in the kind of innovation that I
see from Genuitec,
with its continual efforts to deliver
open-ended Java development capability in the interesting and
cost-effective MyEclipse
subscription model. I'm interested in new offerings like TeamCity, introduced
this month by JetBrains to
provide a flexible but comprehensive foundation for team development
efforts.
I don't dismiss the well-informed comments of someone like Sun
VP/Fellow James Gosling when he expresses
concern that students learning programming aren't getting training
that's either effective or relevant to their future work. On the other
hand, I've never seen anyone learn programming without a good reason—that is, without a problem they really wanted to solve that a cleverly
written program would solve.
I like to think of programming environments as workshops that enable
learning, not as classrooms that impose it.
Tell me what your favorite tools have let you learn at peter_coffee@ziffdavis.com.
Check out eWEEK.com's for the latest news, reviews and analysis in programming environments and developer tools.
Commentary 
