July 04, 2003

Darwinism and Standards

In an interview with Computergram, which is bound to get lots of press coverage and comment in the next few days, Cape Clear's CEO Annrai O'Toole makes the following interesting comments:

Web Services Standards Armies Should Fight to the Death

The standards war that is being played out in the crucial area of web services choreography between the rival Business Process Execution Language (BPEL) and Web Services Choreography Interface (WSCI) camps should be fought until only one is left standing, according to the CEO of web services integration vendor Cape Clear Software Inc, Annrai O'Toole.

O'Toole, who was co-founder and CTO of Iona before he started Cape Clear, said in an interview with ComputerWire: "It is the simplest thing in the world to get a compromise in the area of standards. But you end up with a junk spec. The worst outcome is some silly compromise - just stapling them together. We need one or the other, not both."


While I think the ease of achieving a compromise is slightly (!) overstated, it is certainly the case that it's a Highlander end game here - "There can be only one" [standard] That's what customers and users actually want, and whether that is achieved by compromise or extinction is largely up to the players involved.

Ultimately, you can't fight market forces - if a specification is achieving traction over another, then it *is* the de-facto "official" standard bearer. That's exactly what happened with SOAP over ebXML, for example. That's also what is happening with BPEL too.

Entry categories: Standards Web Services
Posted by Jorgen Thelin at July 4, 2003 11:25 AM - [PermaLink]
Traceback List
I'm interested to know why you equate SOAP with ebXML - isn't that like equating Java with .NET? ebXML seems to me to have embraced SOAP as an option. Posted by: Simon Phipps on July 4, 2003 06:24 PM
Hi Simon, Personally I view 5-sixths of ebXML as largely off the wall - "spontaneous business relationships" will *NEVER* happen in practice, IMHO. What's really left then is just the messaging part, which as you mention is based on SOAP. IMHO the reason ebXML is getting so little traction is *precisely* because of this 5-sixth overhead - nobody is going to buy in to ebXML to only use one sixth of it. It seems to me that people are starting from the other end (SOAP), and adding the missing pieces (security, reliable messaging, conversation, etc) in layers. Sure, they will probably end up with something very close to the ebXML messaging component, but in a manner that can be composed and tailored to their usage scenarios rather than in a one-size-fits-all approach. I'll pass on the Java-vs-.NET bait hook, thank you ;-) Posted by: Jorgen Thelin on July 6, 2003 08:32 AM