Why CMIS 1.1 Is Pretty Awesome

Before I get to the meat of this post, I want to start with a confession. I have been a slacker. If you look at my word cloud to the right, you’ll see CMIS as a big piece of the proverbial Pie. Even before it was a public term, I railed for the need for a standard in the Content Management space.

Now that the first update to the Content Management Interoperability Standard (CMIS) has been out for nearly three months, why am I just now blogging about it? Now that  browser binding, retention, holds, and type mutability have been added to CMIS, why am I not proclaiming the wonders of CMIS 1.1 from every rooftop.

I…uh…got busy.

What I want to do today is talk about why this update means everyone should be looking deeper into CMIS and reconsider it for every Content application created. In fact, as much as the need for standards in Content Management existed when I started writing about them, it is even more urgent today.

What is in CMIS 1.1?

There are three big items in the update. They address three different contingents of adopters.

  • Retention and Hold: As Cheryl McKinnon so aptly covered, the addition of Retention and Hold features adds Governance to CMIS. It isn’t Records Management but it IS a way for organizations to manage records remotely.
  • Web Binding: There were always complaints that the REST binding for CMIS wasn’t useful enough to actually use in composite web applications, known as “mash-ups”. If you wanted to pull it into a Google map, it was challenging. The new JSON binding addresses those concerns.
  • Type Mutability: This is the part that really rocks for solution developers. Before, if you developed a good Correspondence Tracking System, you had to still create the content model in each vendor’s repository before you could leverage CMIS. Now, solutions developers can create the content models using CMIS. That makes the creation of an application using nothing but CMIS not only possible, but practical.

There are more changes but those are the most important additions. Developers, solution vendors,and the Governance types all win with CMIS 1.1.

Why CMIS is More Important Than Ever

When this all started, the silos of information we were trying to connect were all located in our data center. Everyone had multiple Content Management vendors running inside of their shops. The dream of one repository to rule them all was slowly being erased from reality because migrating data into a single system is hard.

That is why CMIS was so promising. We could point applications and solutions at one or many repositories and not worry about what happened behind the scenes. It was great.

Except that adoption among developers has been limited.

Meanwhile, the problem in the data center has escalated while new vendors have entered the market. Cloud vendors offer users the easy experience they have been craving. Adoption is accelerating even as you read this.

Making Content Management Easy

When talking about making things easy, it isn’t enough to simplify things for those that use the Content Management Systems (CMS) to get things done. The owners of both the systems and the information need help as well. When one realizes that a multitude of systems isn’t “crazy” but the new normal, how does one survive?

You survive by finding a way to centrally manage information, even if you are using a Federated tool. Note that I said centrally manage, not store. You develop front-end solutions that not only access the primary repository for that business need, but one that can search all the other information repositories for information.

You use open standards and insist that your vendors support open standards.

With every standard, there are many vendors that support them. In fact, as you look at a vendor, you’ll notice that they will typically support either very few standards or a large number of standards. Support for open standards is a DNA thing, you either “get” it or you don’t.

Do your vendors “get” it?

Why don’t you start by asking how they support CMIS?

3 thoughts on “Why CMIS 1.1 Is Pretty Awesome

  1. Lawrence, I do agree with you on the importance of CMIS and it is a pity that is quite poorly supported by many major vendors. As a Foundational Sponsor of OASIS – https://www.oasis-open.org – we support CMIS 1.1. both as an archive and a client. As it happens there is little benefit because customers don’t really aski for it and when WE ask for it, either the customer’s software is downlevel without CMIS or lacks the right level of CMIS. We are always pointed towards native APIs.

    While type mutability adds a lot of flexibility, I would not claim that one can now develop a complete application using nothing but CMIS. CMIS is for storing and retrieving content blobs, no more. Great that there is a standard way to allow application developers to access archives in a version independent way using their own metadata definitions.

    I would say that the expectations you put into CMIS can’t be fulfilled. CMIS doesn’t make you or your application vendor independent. One problem is that most standards are too little, too late and too complex and therefore there is still a lot of custom interfacing. Most vendors do not want to make their system so open that they loose the competitor lockout. But proposing that they use open standards and hoping that the buyer doesn’t check what it means is a common approach.

    Having said this, the higher number of supported standards doesn’t make vendors better vendors. Even if a vendor supports BPMN, he still locks you in with the other 80% of non-standard process definitions. Even if a vendor or application supports CMIS, all it will do is ease integration work. No more.

    And that is good because the integration work is the actual cost and effort lockin. CMIS is just a small piece of the whole content plus context plus process mangement puzzle …


  2. Paul Wilson says:

    An interesting post.

    Recently we (Murdoch University in Western Australia) wanted to join Moodle to SharePoint 2010.The first stop was, of course, CMIS. When that didn’t work we went to hand written code.

    The problem was, simply, that Moodle has a partial CMIS 1.0 implementation; SharePoint has a partial CMIS 1.0 implementation, and the partials didn’t match so we couldn’t actually get anything to work.

    From a user perspective, CMIS is a great idea, but from the software vendor not so much. It can easily be seen to allow an exit path from a product which may not provide all that was promised or is required. Reading the broken promises from Microsoft about updates to the CMIS code in SharePoint certainly makes you think that way. Will CMIS 1.1 be treated differently? Who knows, but it would be nice to think so.

    Regards, Paul Wilson


    • Paul, the vendors develop to the money. They listen to license deals and to partners who develop solutions that drive license deals. In the cloud, the same thing, Annual Recurring Revenue.

      If there is a vendor, like Microsoft and Moodle, with CMIS implementations that don’t work, we need to call them on it. If Alfresco’s CMIS 1.1 support that was recently released in Community doesn’t work, call US on it.

      I’m a big believer in the fact that if you make it hard for a customer to leave you, they’ll stay gone. If you are open and give them options, they may decide the grass isn’t greener and return. People have a long memory in this industry.


Comments are closed.