My article on the register.
MS and Oracle’s big dev tools – who needs ’em?
Published by jeffvroom
I'm a software architect with 35 years of diverse experience. Initially in low-level graphics software including work on the X window system for the GS1000. Lead developer at AVS, then chief architect of AVS/Express. Senior architect and then chief architect at ATG. CTO at Tribal Fusion, Principal Scientist at Adobe for 3.5 years. More than 10 years as a software architecture consultant, performing tech due diligence and writing challenging code. View all posts by jeffvroom
Published
Jeff Vroom's Blog 
Good article Jeff,
couldn’t agree more.
As responsible for dev tools at a company in the auto industry, I hear the developers asking for more flexibility and lightweight tools. While the big irons comes to demo bloatware ….. Eclipse and the other open source tools is rapidly replacing the expensive suites that we have used before.
We want to do the integrations ourselves so that we can pick and choose tool setups. If we choose Visual Studio, then we more or less must use the whole M$ setup. The same with Rational/IBM.
Thoughts on the Gosling exit from Oracle ?
I’m sure the culture over there has changed pretty dramatically so not too surprised. Probably losing Josh Bloch to google last year was a bigger loss for Java though. Of course neither of them really kept the Java community on the right track from my perspective. Decisions like EJB1 and 2 were pretty obviously terrible to me from the beginning. At ATG, we tried really hard to not implement those parts of J2EE but were contractually obligated to by Sun which really cost us in terms of wasted engineering resources. And it took a decade for them to finally back a somewhat reasonable persistence standard. I don’t blame the engineers specifically since from one or two meetings with Sun execs that I went to, it was clear it was a lack of management vision that was at the core of the problem. Specifically, I explained to them in detail why their app server architecture was terrible and needed to be tossed (anyone remember “kiva->iplanet?”). They honestly acted like they didn’t think the architecture was important. A decade later they started from scratch again with Glassfish. Given how much money gets invested in enterprise software each year, how costly refactoring/rewriting this stuff is down the road, how long companies are faced to live with these bad decisions, these are all pretty costly moves from the industry’s perspective. Right now my magic8ball is saying things will only get worse with oracle.