Blogs

I am an incurable optimist. I believe in the power of technology and that with software anything is possible. So it is understandable that on a panel on CMIS at AIIM the other week, Microsoft program manager Ethan Gur-esh described my vision of what might happen with CMIS as grandiose. If you are not familiar with CMIS, it is the Content Management Interoperability Services, which has just been ratified by the member companies of OASIS. It is in effect, an SQL for Content Management. In other words, a set of protocols, interfaces and query language definitions that allow a program to be written to any content management system.

 

On the panel, I described how I believe CMIS can transform the ECM industry, allow for significant growth and spawn whole new companies and markets. Yep, that sounds pretty grandiose! However, on the same panel Ethan announced that Microsoft will be providing support for CMIS 1.0 as an extension to SharePoint 2010 in June. Al Brown, from the more conservative IBM / FileNet, didn't provide forward looking statements. But with the DeveloperWorks implementation available and the amount of participation that both the FileNet and Lotus divisions have provided to CMIS, the hints couldn't be stronger. Ryan McVeigh represented not just Oracle and his Web Center organization, which manage Oracle's many portals, but also the content management group including Stellent. Again the support was strong and all these organizations, Alfresco and many more have already created and tested interoperability between many different CMIS systems. The moderator, Mike Mahon, is from Zia Consulting who has created mobile applications for the iPhone and Android using CMIS.

 

In short, all the major vendors are all behind CMIS and it will soon be possible for developers to build applications that can run on any repository. Also, the level of functionality in CMIS is quite rich even if it doesn't cover all the functionality of an ECM system. It's enough to build some seriously interesting applications. Are there any precedents for this level of standardization, interoperability and functionality? Many standards, such as various workflow, business process, web services and XML standards, have not been able to achieve these levels of compatibility.

 

Really the last time that you had anything like this was with SQL and the database industry. (I am old enough to have been around and have a management position in the database industry then.) Granted, SQL was done in two steps with SQL 89 and SQL 92 and the level of completeness was probably a bit higher. CMIS is also lacking some of the data definition capabilities of SQL (and JSR-283 Java Content Repository 2). However, with two protocol bindings, a complete (and simple) domain and object model, a whole set of capabilities around content handling and versioning, and a security model, CMIS is broad and rich enough to create many applications that have been recreated over and over again. Now however, these programs don't need to be tied to any particular vendor's repository - other than data type definitions.

 

Starting more or less around the time of SQL 92, whole new categories of companies were created with databases at their core. A lot of the client-server generation of companies were started that depended upon the database, such Siebel and PeopleSoft. SAP moved from a mainframe application to a platform and database independent client-server app and grew at an explosive rate. At Documentum, where I was a founder in 1990, we initially built our document management system on a Sybase database and quickly ported to Oracle. We did well and ultimately went public and Oracle did even better than we did on the back of selling databases that supported Documentum. Coming as it did after the 1991 recession, SQL was well aligned to take advantage of new investment in technology that often follows a severe downturn in the economy. Over the period of that decade, the database market grew tenfold and much more since then.

 

Companies that have already built or are thinking about building solutions on top of content management systems can now think about the problem very differently. Why? Because that is what happened with databases with SQL. Rather than think about not just the technical aspects of building the applications, new solutions developers do not need to think in the context of any particular vendor. This gives them the freedom to concentrate on the value that they are adding for the customer without needing to think about how their solution fits into the ecosystem of a specific vendor. Solutions providers knowing that they will have the whole of the $4+ billion ecosystem rather than just a slice, can more boldly invest in solutions that can go into a lot more customers. This will reduce risk of market concentration and raise the level of investment from internally generated cash and, pretty soon, venture capitalists.

 

A lot of money has been made on custom solutions sometimes costing 10 times the original cost of ECM licenses. Much of that money will be channeled into reusable packages of applications designed to work on multiple repositories. New solutions that integrate collaboration and social networking capabilities will also be content rich and need the functionality provided by CMIS. If you were to design an API for content applications in the Cloud, it would probably look a lot like the RESTful protocol of CMIS, so expect Cloud applications and content-oriented SaaS offerings to take advantage of CMIS. This could be the creation of companies the size of those client-server companies and the web-based companies that followed those.

 

The standardization of platforms heralds the opening of markets that rely on those platforms. And platforms, once they are standardized tend to stick around for a long time and grow. Think Unix (to Linux), SQL and Java. Now think CMIS. Think Big, Hairy and Audacious!

 

I told you I was an incurable optimist.