Burton Catalyst

I was at the Burton Group's Catalyst conference, in San Diego on Wednesday. Very impressive, they had something like 1600 attendees, architects and the like. I was there participating on a panel discussion moderated by Anne Thomas Manes, with a number of VIPs from other companies:

  • Bill Roth, VP Product Marketing, BEA
  • Jamie Thomas, VP WebSphere Development, IBM
  • Rick Schultz, VP Application Server Marketing, Oracle
  • Kaj Van de Loo, VP, Market Development Engineering, SAP
  • David Bryant, Senior Director of Marketing, Application Platforms Products, SUN

A solid set of titles, all the way around!

The discussion moved around quite a bit, but covered open-source support within the various vendor platforms; using open source to engender a developer community; the value of being "open" versus interoperable; application platform strategy; model driven development; how commercial platforms distinguish themselves from "good enough" alternatives, and other topics.

Anne made a couple of provocative statements, including:

  • interoperability trumps portability. Businesses get little value from Portability, and it's not possible anyway (even with J2EE).
  • "super platforms" that combine and integrate many facets of an app platform are a "disruptive innovation" in app platforms (ref The Innovator's Dilemma)
  • the key disruption in superplatforms is the cohesiveness of the piece-parts. So, integration wins over "best of breed"
  • The Windows Server System is the leading example of a cohesive superplatform.

All of which I agree with! What I found interesting is that there were no strong objections to these, especially the first statement, about interop.

Some of the participants on the panel also made some interesting statements.

  • J2EE (the spec) is essentially designed to impede innovation by vendors - from Bill Roth. I agree but would expect this to be provocative.
  • Portability among Java-based app servers isn't really possible anyway. (I forget which of the vendors stated this. Again, I agree!)

For my part, I stated that I thought J2EE was like CORBA and DCE - it moves just too darn slow to be sustainable in the long term. Therefore vendors look outside the J2EE spec for advances, to things like struts, ant, Eclipse, or home-grown things like Oracle's ADF, or Sun's framework, or WebLogic's Workshop framework, or IBM's WebSphere extensions, or SAP's extensions. All of these extensions lead directly back to the lack of portability, mentioned by several people.

Anyway, it was a good discussion, I hope valuable to the audience.

-Dino