So it was surprising to hear at EMC’s recent Documentum conference in Prague that the content management juggernaut was spearheading a new OASIS standard, content management interoperability services (CMIS), and even more surprising to hear it was collaborating on it with Microsoft, IBM FileNet, Open Text, Oracle, Alfresco and even SAP.
EMC’s vice president of marketing for content management, Whitney Tidmarsh, promised CMIS would be “not unlike the SQL standard for the database industry”, and would “dramatically lower the total cost of ownership for multiple systems”. CMIS, she claimed, represents a “maturing of the market” and “is a tipping point in content management”.
Documentum’s CTO and founding engineer Razmik Abnous was equally enthusiastic about the standard, saying it would greatly impact independent software vendors (ISVs) “who don’t want to write apps for a platform, but for the market”.
“This is very big for us, it’s not just a paper exercise. Without integration, the industry would not grow in this way,” he said.
This is reinforced by one of the founders of Documentum, and the CTO of Alfresco, John
Reality check
However, some have reservations about the initiative. Comparisons of CMIS to SQL are erroneous says David Nuescheler, CTO of web content management firm Day Software and specification lead on
the Java Content Repository (JCR) standards JSR170 and JSR283, sometimes considered ‘failed’ attempts, depending on which ECM vendor you talk to.
“SQL is a query language and is not good for interoperability, as anybody who has had to port a database would know,” he says. “It was not built as an interoperability standard but as an infrastructure standard.”
“An infrastructure standard is a stated non-goal of the CMIS, and that kind of standard can drive consolidation – what SQL did for the database industry was drive everyone except three players [Oracle, IBM, and Microsoft] out of business, which was great for interoperability.”
CMIS, explains Nuescheler, in contrast to JCR, deals only with the concepts of ‘document’ and ‘folder’. “There are big omissions, a lot of holes I can see in it – web content for us, in particular. It’s a very focused specification which may help it be successful, although it will not help programmers all that much. They still need an API.”
The reason for the sudden enthusiasm for standards among the large vendors is, says Nuescheler, “a very good question.”
“While it’s a move in the right direction there are [already] enough standards out there, they just need to support them. Interoperability is not the fault of the standards, it’s the vendors’ [lack of] will to interoperate,” he suggests.
“In my experience [content management] is one of the most difficult things to work with in terms of gaining consensus. It is hard for them to go in and modify an existing product with a large install base. In the early days, when we were working on JCR, several vendors came up to me and said, ‘vendor lock-in is a good thing, why are we doing this?’”
What CMIS might achieve, Nuescheler hopes, is to give vendors a new momentum. “They’re excited about this,” he says.
But for their customers, “the point where we see if it actually works remains to be seen. So far, technical discussion in the committee hasn’t started yet.” Buyers, meanwhile, “should go with a vendor that has at least a strategy around standards.”