Gartner Blog Network


CMIS 1.0 May Be 3.0 Years Too Late

by Toby Bell  |  January 13, 2011  |  11 Comments

It might be hard to determine given evidence of other interests in my blog, but my primary research coverage at Gartner is focused on enterprise content management. For those of you not interested in this (and you outnumber those that do by a nearly immeasurable factor) I will keep the technical micro-humdrummery surrounding standards to a minimum here. The details (available on gartner.com) warrant some attention, though, because many efforts to create and then enforce standards fail – and the one I reference in the title of this posting should stand a better chance at adoption than many. How, you may wonder, did I come to the conclusion that a new enterprise content management standard (circa 2010) might be too late to succeed? It came to me as ideas often do: while driving in blinding rain on the wrong side of the road.

Let me explain. I travel often enough to encounter all kinds of differences between people, places, and things. Yet, in many countries there are similarities even among the differences. So, for example, when traveling in the UK, Bermuda, and Australia I can expect to hop into any automobile and drive it from the right side forward seat on the left side of the road. But I’ll pay for traffic violations with different currency even if a Queen is common among them. Elsewhere, the opposite is often true: left side driver on right side of road. However, a few weeks ago I rented a car in St. John. There, I drove on the left from the left. And, it seemed to me that others were probably doing the same (though as I said, it was raining).

So, the first rule of standards might seem to be that they are not often global – whether in a geographic or even a company-wide sense. And, since Enterprise Content Management is as much a contradiction in terms as Naked Mole Rat…

ECM NMR
not cohesive – rarely a single system for all content across enterprise not actually ‘naked’ but lacking fur (and/or our notion of modesty)
not conclusive – content means different things to different sponsors and users not a mole at all – a separate species
not controlled – management implies governance and/or lifecycle control not a rat at all – a separate species

… It stands to reason that a standard which allows information to be better managed and accessed would be a good thing. However, to this point every enterprise content architecture is unique – and typically business leadership could detail “failure measures” more often than success ones. Maybe a success measure is allowing everyone a common access protocol to a virtual repository stitched together by a standard. But I also refer back to the first rule of standards and now add more detail: not only are standards rarely global; standards can also be deadly. Am I saying that CMIS 1.0 will kill people? Aloud?

Not yet. But in Contentville there are drivers (sponsors) and there are pedestrians (users) and protections are generally better for the former. And, without repeated warning, any changes to the environment may cause sponsor/user collisions, especially if users fail to >> LOOK RIGHT >> (as clearly indicated on the footpath) prior to crossing the road. In the real world, 12,567 pedestrians from countries with opposite traffic rules die each year in the UK, Bermuda, and Australia as a consequence of the absence of global standards. Amazingly, there are also 12,567 people in those countries who annually earn a living by painting warning messages on streets which are ignored by foolhardy tourists. So, effectively it’s a wash. And I’ve made a sorry mess of the analogy, too.

More interestingly, many of those people paid by certain countries to paint reminders on sidewalks reminding tourists of standards differences also continue to rely on non-standard coinage for many day-to-day transactions.  Even as inflation limited their spending power, their governments would produce more and more varieties on the same theme. It seems incredible that even as population trends indicate a small decline in the birthrate benchmark figures of industrialized nations – thus showing consideration for our overburdened planet – their pocket money has blossomed to the extent that at the end of any given day they carry the equivalent weight of a full-term newborn in their trousers.

Some of the coins have “jingled together” in pockets long enough to have spawned a second generation of tiny coinlets which numismatists once wrongly guessed were actually the pets of the larger denominations. My theory on a related subject: it’s the weight distribution of coinage in Europe which is causing the Earth to spin crazily on its axis, further stimulating global climate change. Furthermore, we may have to put both ‘belts and braces‘ on some of the smaller countries to keep them from slipping right off into space. So, the hope of all people-kind now rests on the euro as a standards superhero.

With the advent of the euro, a common measurement of value now exists for stuff across most of Europe. This represents more of a problem for some countries than others. The euro bills and coins didn’t yet exist even as committees populated by caring representatives from all nations struggled to assign meaningful figures (for solid basis of comparison) to all goods and services.  The cost of souvenirs is a good indicator – how much should a bullfight-themed folding pocket fan from Madrid cost by comparison to one which touts the Grand Canal in Venice?  Folding Fan Fights alone have cost at least a year of factional squabbling in the souvenir subcommittee on value-based pricing. Unfair global currency valuations, bailouts, and other inequities now plague the euro and cast significant doubt on its survival. If the euro falls, I presume more fragile standards like CMIS might too.

In terms of actual usage, establishment of new universal standards tend to yield lower enthusiasm than they might have during some ‘golden age’ of the past. Especially if the standard was actually gold, as I heard it may have been at some point in history. Because now… guess what?  Enterprise content management technology is the ‘gold’ of the new information economyTM. A YouTube video used for sales training or product promotion – it’s content. A tweet related to project team meeting location – more content. An iPad application that presents ideas as text or images or full-motion media – again, content. A contract or presentation or invoice or sales order – all content. And, considering the difficulty of navigating the 4-15 different repositories most enterprises maintain for contribution and consumption of said content by people, something that allows all those repositories and users to interact more seamlessly would be of huge importance. And, perhaps as more content is contained within cloud hosts some kind of common logic about the containers and their access that creates a unifying user experience for those dependent on a hybrid content architecture (some stuff on premise/other stuff in the cloud) would prevail and be good. Such is the unique value proposition of CMIS 1.0.

Sadly, people tend to reveal differences and then revel in them. So, it’s likely that sponsors of some systems won’t want to provide access to lots of users or restructure the underlying repository or change the user experience in any way. Some vendors will use the upgrade path to CMIS-enablement as an opportunity to charge for upgrades and services. Others will embrace it as a way to keep their clunky technology relevant and forestall displacement. Moreover, It’s a developer standard in the era of SharePoint solutions and repositories out of reach like Facebook, Twitter, YouTube, LinkedIn, and Flickr. It’s a document management standard in an era of agile information. For CMIS to take a rightful place in the pantheon of de-discombobulating technology tools, it will have to attach itself to a high level strategy for designing content architectures for ideal user experiences as well as identify better ways of ‘relationalizing’ information of many kinds across many places. And, that might take more time than the market itself allows. Even as I write this, enterprises are adopting another approach to reconciling the content management mess. They’re adopting SharePoint, migrating content to it from network drives (and sometimes competing ECM products), and enforcing it as the universal content infrastructure. Whatever they use for governed content (the repository of record) already has a SharePoint connecter. And, now they’ll use SharePoint as the client for all their content.

So, a market standard emerges to trump any attempts by a technical one. And now, they’re waiting to use this platform as consumers do their iPhones, iPads, or Android devices – as the landing place for content apps downloaded (for dimes, dollars, euros) from an apps store. CMIS is designed to help lots of ECM applications work together as one to provision users better. But enterprises effectively started to formulate a content strategy when SharePoint 2007 emerged. CMIS allows technologists to hook up bunches of repositories that didn’t work together well or present information in the same way to users. In effect, it’s the savior for tired technology and rationalizing past bad behavior by the buyers of it. What CMIS could become instead is the governing review and certification approach for content apps – and how they play well with platforms and other information. Or maybe there’s an ever better path to success.

My Gartner colleague – Larry Cannell – suggested as much long before CMIS 1.0 was ratified: “Compliance with CMIS mandates a content system provides a fairly minimal set of capabilities, primarily support for navigating folders and opening documents. However, the optional features in CMIS could provide the basis for a significantly richer exploration of repository-based content by using metadata. For example, instead of being limited to a simple folder hierarchy, imagine content being explored by type of file, by date, or by custom fields that are meaningful to the business such as sales account, plant code, or engineering release. Content wouldn’t just be stored in one directory but could be found via multiple paths using this metadata. For example, a design specification might be found by exploring multiple paths such as product segments, regions, or suppliers. Content management vendors have been demonstrating these opportunities in CMIS for months.”

AHA! As I’ve said in earlier posts, metadata is the plastic of the semantic Web. Content applications are the gold. Combine the two and what do you get? Exactly. DITA. An XML standard for describing information across an enterprise (and even an industry). So, if we could get the CMIS folks to work together with the DITA team…. We end up with SCIMITAD: a sword that cuts content into meaningful components for human consumption.

What do you think? Are standards for content management systems and interactions more or less meaningful now for your enterprise (or platform as a vendor) than a few years ago?

Additional Resources

View Free, Relevant Gartner Research

Gartner's research helps you cut through the complexity and deliver the knowledge you need to make the right decisions quickly, and with confidence.

Read Free Gartner Research

Category: somewhat-serious  

Tags: cmis  content-management  dita  ecm  nmr  standards  xml  

Toby Bell
Former Research VP
8 years at Gartner
25 years IT industry

Toby Bell is a former vice president in Gartner Research, responsible for a range of document and content management technologies for the High Performance Workplace group. Some key areas of coverage include vendors and trends in the enterprise content management (ECM) marketplace… Read Full Bio


Thoughts on CMIS 1.0 May Be 3.0 Years Too Late


  1. […] This post was mentioned on Twitter by Suvish Viswanathan. Suvish Viswanathan said: CMIS 1.0 May Be 3.0 Years Too Late: It might be hard to determine given evidence of other interests in my blog, … http://bit.ly/dXckcG […]

  2. Paul Hampton says:

    Hi Toby,

    As always a very thought provoking blog entry.

    I agree that we have not seen the ubiquitous ECM solution and organisations will continue to use a range of ECM (Alfresco, Documentum, IBM Filenet, SharePoint, etc.) depending on departmental or functional need. And due to some of the limitations in SharePoint I do not see it taking that place any time soon.

    But surly the benefit of CMIS is in this environment where the end user needs to use multiple systems. We are seeing new applications coming out that are CMIS aware. These allow end users to connect to any CMIS compliant repository using that tool. Removing the need to learn a number of different interfaces.

    From an application vendor point of view building a new CMIS client means only having to code to one set of API’s rather than a range of proprietary API’s. This will lead to new and innovative application that can access content irrespective of how it is managed.

    CMIS 1.0 too late? I think not, but definitely long over due.

    Paul Hampton
    (Disclaimer: I work for Alfresco, an Open Source ECM solution)

  3. Toby Bell says:

    Thanks, Paul. Has DITA adoption become a critical consideration for Alfresco? Does a combination of standards for authoring, storing, using, and classifying componentized content make sense yet?

  4. Paul Hampton says:

    Hi Toby,

    DITA and XML are very important standards for Alfresco and the industry as a whole. We use DITA within Alfresco to author and publish our own technical documentation. Storing XML ‘topics’ inside Alfresco that can be published a number of different ways (e.g. print, PDF, online help, etc.). But DITA is typically still consigned to the realm of technical documentation.

    But these standards are different from CMIS. CMIS addresses how systems interact. Allowing applications to be built independent of the underlying content management system.

    CMIS will allow portability of these applications. Which should increase the market potential for new content management applications, which in turn will allow developers to invest more time and money building better clients. All of this will drive new innovation within the CMS market.

    A much over used phrase, but think of CMIS as the SQL for Content Management.

    Paul

  5. Toby Bell says:

    Thanks, Paul. I suspect enterprises are ready for a complete standards ‘kit’ that can help their content management systems, the information contained within them, and end-users themselves interact better. I may be dreaming… but I’m not the only one.

  6. Sam Babic says:

    Great article Toby. As always your analogies create a great vicarious experience. I’m now second guessing myself when I cross the street.

    A comment on one of the quotes from Larry Cannell taken from above:

    “Compliance with CMIS mandates a content system provides a fairly minimal set of capabilities, primarily support for navigating folders and opening documents. However, the optional features in CMIS could provide the basis for a significantly richer exploration of repository-based content by using metadata…”

    This comment speaks to the upside of CMIS, but there is a downside (and not so much CMIS but typically of standards in general) especially where the standard has provisions for too many “optional” features.

    Accepting only lowest common denominator features can be useful to reduce the cost of ownership to integrating the various systems within an organization. If an IT department only develops solutions to the lowest common denominator components of CMIS, then there should be little variability in the design of their solution as presented to the end user. What does this translate into? Lowest common denominator. Good for IT but bad for end user experience.

    The end user should be able to take advantage of “optional” features of one particular vendor over another. If CMIS is used to expose these optional features, there is now an increased cost and complexity to the IT group that is developing the solution. Suddenly the “standard” requires that the IT guys understand each vendor system and in some circumstances potentially the inner workings of that system. Similar to how HTML had/has many optional features that vary by browser and version of browser. ODBC is another great example of a standard with optional features that can potentially make the implementation of a solution very non-standard. Is it better than the alternative? Depends. Is there hidden cost? Yes.

    Even if the IT department pulls this off, it may still translate into a bad user experience. Users may be abstracted from the repository but may be wondering why “feature X” works on this document from this repository and not on a document from another repository. A consistent experience across repositories may not be achievable if too many optional features creep their way into a solution. The experience is really a function of (1) How many “optional” features are there, and (2) How “low” is lowest common denominator relative to the most feature rich system.

    There is one other facet of standards too that can also contribute to hidden costs. When a new version of standard comes out you may not be able to take advantage of it across all repositories. If you have lots of repositories, the likelihood that you can upgrade all in lockstep is probably slim and if you could hopefully they all comply with the newer version of the standard. Otherwise, IT must introduce an additional permutation within their solution to accommodate systems that have not yet been upgraded or that only comply with features from an older version of the standard. Suddenly the IT guys are putting a facade layer on top of their already facade layer.

    Having said all that I’m not nearly as negative on the topic as I appear. I’m simply pointing out some potential gotchas and hidden costs that organizations should consider when taking on any IT initiative to align with a standard.

  7. Toby Bell says:

    Great feedback, Sam. Doesn’t all this point toward cloud-hosted repositories and the potential benefits in standards, scale, and support they propose? The complexity of on-premise content architectures (whether pre- or post-CMIS) may be too much to bear for business leadership unless ROI proof points for upgrades are crystal clear.

  8. Hi Toby,

    Great post. Very entertaining and very true.

    Too little, too late?
    Timing is crucial, no doubt, but I would argue that CMIS functionally and from a technology perspective (XML over HTTP) is equivalent to WebDAV. Now, WebDAV was more than three years earlier and every computer/os on this planet has a WebDAV client built in and every ECM vendor and beyond (even MS Exchange) supports WebDAV. Great adoption.
    So why did the ECM community have to reset all their efforts in start from scratch? And did it matter? I don’t think it mattered at all…

    CMIS is great for software vendors like us, since we can use CMIS to interact with the document management legacy systems building connectors without having to understand a cryptic API ported over from the 80s. Does that mean they are all the same and we can use the same CMIS connector for all of them. Of course not, not even close, but it makes our lives a lot easier. On top of that good news for us and all the other CMS vendors based on JCR infrastructure all become automatically CMIS compliant, since the CMIS is a compatible functional subset of JCR.
    Standards in general are written often from the point of view of the Software vendor or the implementer. Which in my mind is the core problem.

    For standards to become useful for the “business users” of the respective technology their interests must be represented in the specification process. Essentially driving a specification not as a lowest common denominator minimal featureset but defining true and honest infrastructure that offers value to their users rather than trying together legacy.

    In my mind all successful infrastructure like HTTP, URI, HTML start out with software vendors wanting to invest and push things forward.
    I think if “web standards” were specified with the same attitude as “enterprise interoperability specs” we would still try to find the common grounds between IE 3 and Netscape 4.

    regards,
    david

  9. Toby Bell says:

    Thanks, David. You’re right: “For standards to become useful for the “business users” of the respective technology their interests must be represented in the specification process.” You put it more succinctly than I. We need to hear from more end-users about the perceived value of CMIS (and a kit of other standards) in their ECM strategy.

  10. […] and whether this new OASIS standard will really matter in the ECM technology space. Titled “CMIS 1.0 May be 3.0 Years Too Late“, he uses an interesting analogy of local driving habits and […]

  11. Toby Bell says:

    Hi Cheryl. I liked your post, too – and encourage others to read it.



Comments are closed

Comments or opinions expressed on this blog are those of the individual contributors only, and do not necessarily represent the views of Gartner, Inc. or its management. Readers may copy and redistribute blog postings on other blogs, or otherwise for private, non-commercial or journalistic purposes, with attribution to Gartner. This content may not be used for any other purposes in any other formats or media. The content on this blog is provided on an "as-is" basis. Gartner shall not be liable for any damages whatsoever arising out of the content or use of this blog.