XML founder takes on WS-splat

Tim Bray, founder of XML, has never been a big fan of the WS-* specs -- he has referred to them in the past as "bloated, opaque, and insanely complex." In a recent Weblog post, he weighs in on the issues raised by Gartner's Daryl Plummer in Optimize regarding the suitability of the Web services stack (mainly WS-splat) in supporting enterprise SOA. (Also reported here in this Weblog; and I dive into deeper analysis here.)

Bray agrees with Plummer on many points, especially regarding the viability of REST (XML over HTTP) as a simplified option for application-to-application interfaces. "When it comes to Web Services, you can choose between simple, reliable, standards-based infrastructure that’s here now and the sprawling, shifting WS-* technology that’s still under construction, mostly by IBM and Microsoft, both well-known champions of simplicity."

However, he does take issue with Plummer's dividing the Web services world into "two camps" -- one for Web services-based SOA for internal enterprise integration, and the other for external interactions using emerging Web technologies. "I’m kind of puzzled," Bray wrote. "It seems to me like the pressures on both inward-facing and outward-facing applications are pretty similar: operation across heterogeneous systems, stringent reporting and auditing requirements, high performance. And Service-Orientation is pretty important out there on the Net, too... the notion that the term 'SOA' is reserved for talking about internal systems seems dangerous and wrong."

Speaking on behalf of his employer, Sun Microsystems, he acknowledges that "Microsoft is making a huge, company-scale bet on the WS-Future; a really remarkable amount of work by Really Smart People went into WCF. The customers that that adopt this stuff will need to work with the Java universe and with our enterprise products, and we’ve said we’ll work with Microsoft to make that happen."

However, speaking his own mind, Bray adds that "I think that we ought to be pouring resources and investment into tooling and developer support around simple XML/HTTP/REST technologies. You know, the standardized ones that work today."