Comments on: CMIS, APP, Zen-SOAP and WS-KitchenSink: some data points

By: William Vambenepe’s blog » Blog Archive » State modeling: party over, go home now.

Wed, 24 Sep 2008 05:27:35 +0000

[...] advance?) the 5 years-old debate between the WSRF, HTTP-only and “no convention” (e.g. Zen-SOAP as used in CMIS) approaches to interacting with stateful resources over the Web. If you were anywhere near OGF [...]

By: Craig Randall

Craig Randall — Sat, 20 Sep 2008 05:57:22 +0000

I’m glad that not being part of the content management space professionally didn’t deter you from posting your unique perspective on CMIS. Specifying two protocol bindings for a common domain model is expressly about increasing interop’s accessibility (e.g. SOAP for system-to-system interactions, REST/Atom for system-to-application interactions). The OASIS CMIS TC convenes next week, and there is much work to be done to get the proposed specification from its current 0.5 version into a ratified v1.0. Your attention and participation in this proposed standard is appreciated.

By: James Urquhart

James Urquhart — Sat, 13 Sep 2008 05:35:30 +0000

William,

Very cool observations. I’ve argued for some time that WS-* lost its way somewhere along the line, and that simplified messages could do much of the same work with fewer bytes and/or operations. CMIS is important for many reasons that have nothing to do with specific protocol standards efforts, but its good to see someone dig into the implications of a) having both SOAP and REST definitions, and b) the decisions by this team of engineers to forgo adornment for simplicity.

Thanks for this.

James

By: Stefan Tilkov

Stefan Tilkov — Thu, 11 Sep 2008 16:30:35 +0000

Is the (somewhat dubious) REST mapping really longer? It seems to include the full Atom syndication schema as well as the schema for the CMIS-specific stuff. My guess is the SOAP mapping would have been the longer one had they chosen to put it first.