Original URL: http://www.theregister.co.uk/2008/10/02/fielding_rest_cmis/

IBM, EMC, Microsoft blasted for 'REST rip-off'

Not just 'idiotic' - 'misleading' too

By Gavin Clarke

Posted in Software, 2nd October 2008 23:07 GMT

IBM, EMC, Microsoft and others have been blasted by the father of Representational State Transfer (REST) for making "misleading" and "idiotic" claims about a proposed specification for applications to talk to different vendors' content management systems (CMS).

Picking on last month's over-hyped Content Management Interoperability Services (CMIS) Specification announcement, Roy Fielding has said he's getting tired of big companies making "idiotic claims about REST and their so-called RESTful architectures".

Fielding appears to feel CMIS has been mischaracterized as a RESTful protocol while it's been tailored to accessing data repositories of legacy document management systems - not the hallmark of REST.

"The only similarity between CMIS and REST is that they both have four-letter acronyms," Fielding said.

Fielding also slammed overzealous IBM and EMC marketing drones for deliberately misleading people by calling CMIS a standard when it's still a work in progress. CMIS was only last month submitted to OASIS for ratification as a standard. Members of the fourth estate also got it in the neck for reporting on CMIS without being able to tell the difference between a standard and a "standards effort".

Fielding, an Apache Software Foundation co-founder, defined REST in his seminal 2000 University of California dissertation Architectural Styles and the Design of Network-based Software Architectures. REST describes an architecture for a system where users pass through different pages on the web. It relies on standards such as HTTP, URL, XML, jpeg and so on.

The crux of Fielding's problem with CMIS from the architectural perspective seems to be that CMIS is geared to a very specific use: accessing data and folders in vendors proprietary databases while REST is considered a general web architecture using open web standards.

Following on from this, CMIS pays just lips service to REST by employing a RESTlike protocol called AtomPub, which he seems to feel has been added for "marketing reasons" along with WSDL.

Fielding called CMIS a Web Services interface that's a "thin veneer" on RDBMS-based data repositories - little surprise given those backing the specification are database, CMS and business application vendors. CMIS is a "classic example of what happens when a control-oriented interface is slapped onto an HTTP-based protocol instead of redesigning the interface to be data-oriented," he said.

According to Fielding REST was a buzzword used to announced CMIS.

Fielding appears to believe CMIS has more in common with the family of Web Services specs Microsoft and IBM developed and have driven, which have been affixed with the WS* moniker.

"It [CMIS] should be "renamed WS-DMS and tossed on the same pile of other specs from that genre" Fielding said.®