Joho the Blog
|
|
|
May 29, 2003
This panel consisted of three people implementing open source content management systems, one trying to figure out what to do, a very smart guy from the W3C and, um, me. Jennifer Lynch (U. of Missouri), has been implementing a Typo3 system. She and a colleague have built a substantial system in just a few months for a total cost of $5,000. Marc Lavallee has been converting Boston.com to Zope. Twelve people have spent a total of 14 months (including a four month RFP process) building a system that will go live in ten days. He figures it is about a million dollar project, all costs included, a figure substantially lower than what it would have been had he been using commercial, proprietary software. Hal Roberts from the Berkman Center has been installing a WebGUI system, which entails importing 175,000 static pages. He chose WebGUI in part because of its object orientation and because it lets you edit pages on screen, not in a separate editing mode. Sam Quigley, of the Harvard Art Museums, is trying to figure out how to figure out which system to use. This raised the question of the value of consultants. I said I thought they could be helpful. Hal replied that it's continuing expertise that ought to be brought in house. The audience had diverse opinions. I personally don't disagree with Hal; in-house tech expertise can be crucial, especially if the project is big enough. But a consultant who keeps up with the field for a living can help match the application needs to the right application faster and better. (No, I do not consult in this field.) Then Dave Winer sparked controversy — shocking, I tell you! — by saying that it's like the early days of word processing when everything was hard and expensive. It shouldn't be as technical as it is. It really should be a $200 solution, he said, that does the 80% of what actually needs to be done. Hmm. I don't think the users on the panel could get what they need in that 80%. They're not looking for a desktop application like word processing. To them, CMS is a system, and it does something complex that will only get more complex. It manages documents and document fragments. It provides versioning. It handles permissions. It moves stuff through workflows. It worries about archiving and records management. It automatically lays out pages. It provides editing tools for content and for styles. It serves up personalized pages. It tracks hits. It enables cash transactions. It plops ads onto pages based on who's seeing them and accounts for every view and click. It integrates with the rest of the office software environment. And if it doesn't do all those things now, that's where it's headed. We got a demonstration of why CMS will remain complex software. Someone in the audience asked if it'd make sense for his small college to get together with a bunch of other small colleges and come up with a set of app requirements so that they could share the cost of customization. The general sponse was: "It sounds like a good idea, but... " For example, Hal and Marc both said that their own installations were unique. Someone in the audience agreed. I pointed out that this reminded me of the SGML wars of the '80s when entire industries tried to build a shared DTD. It turns out that everyone's needs and vocabularies are different enough that trying to produce a common spec is extraordinarily difficult. Now, it certainly can get easier. I spoke afterwards with Bob Doyle of CMS Review who is trying to come up with CMSML, a way of describing CMS features that would work for all content management systems. Bob knows that there isn't one ideal and perfect way of doing it, but believes that you could at least make it easier for customers and users to compare systems. CMS is inherently tough and complex. Implementing a CMS system is always going to require someone with strong skills because it touches the way an organization thinks about and handles documents. It will always require looking at document processes, the social structure, and the power relationships in an organization. It requires understanding the legacy "document schema" and looking ahead to the near-term and long-term futures. CMS will resist commoditization for as long as I can see. Which leads me to conclude: Did I misunderstand Dave's comment? Do we disagree about the definition of a CMS? Or do we just disagree? Unfortunately, I'm about to leave for Chicago - the Digital Genres Conference - so I'll be out of contact for the rest of the day... Posted
by D. Weinberger at May 29, 2003 02:29 PM
|
Comments
I think that Dave thinks of Manila as offering a complete CMS solution. For some users that's true, for others it's not.
One the other hand, I think there's a big question as to whether some big CMS investments are really necessary. People tend to take "requirements" (for workflow, auto-layout, etc.) as written in stone, when the question should be "what's it worth to you, in development cost, maintenance, support/training, etc.?". What are the financials of Boston.com, I wonder?
Posted by: BillSeitz | May 29, 2003 07:49 PM
I think you just disagree with Dave. He's got vested interests (both money and ego) in saying that a one-size-fits-all CMS will work for anyone. But I've never seen that work. Every so-called off the shelf CMS I've ever seen required extensive (and expensive) customization to make it fit the needs of the customer. I've never understood why a company would spend a million or five dollars on Vignette when they would end up doing at least as much programming and customization to get it to fit existing corporate culture as writing something from scratch in PHP or Perl or Python and MySQL or PostgreSQL or Oracle. But I've seen it happen time and time again. Those off the shelf CMS companies must have some powerful sales juju. Or maybe they just learned a lesson from off the shelf packages like PeopleSoft and SAP that purport to enable you to run your company, so long as you run it the way they expect you to.
I built my first custom-crafted CMS in about 1995 to help my group maintain the web site of a corporate research lab, and my most recent one to help me maintain a couple of blogs. Every time I've tried to use one of the off-the-shelf solutions, I've run away screaming, back to the shelter of crafting my own.
Posted by: ralph | May 30, 2003 01:19 AM
Great post David.
I think you shed some light on why it is difficult for the bigger companies to make all their content Google accessible. The content has to live in the database to provide the CMS functions you describe and Google can not crawl the database.
Its great to go beyond - blogging good, big companies bad.
Posted by: Mike Sanders | May 30, 2003 09:00 AM
The permalink for this post doesn't seem to be working...
Posted by: Donna Wentworth | May 30, 2003 10:37 AM
>> CMS is inherently tough and complex.
More precisely, CMS is inherently undefined. To users it is What I (Think I) Need. To vendors it is What I've Got (That is Exactly What (You Think) You Need). A 21st century cover of Emerson(?)'s Blind Hindoos Musing on the Nature of Elephant.
80% of a universal CMS solution is available for free -- a DOM capable browser, XML, XSL, XSLT from a number of sources, a text editor for composition, email for one-to-one communication and audit trail (add $1/month for spam filtering). Just convert your knowledge to text and apply the appropriate tools.
The path to a doable CMS diverges based on the answer to this question -- is CMS something that I do or something that happens for me?
Posted by: Wray Cummings | May 30, 2003 11:14 AM
I have the pleasure of working with Jennifer Lynch and her co-"webdivas" at the University of Missouri-Rolla. What they've done with the site by implementing Typo3 is amazing. Thanks for the props to them.
Posted by: Andrew Careaga | June 5, 2003 04:03 PM
In both my comments on the panel discussion and my presentation, I stressed the major difficulties (the most difficult being the political aspects rather than the technical ones) of incorporating a CMS into an academic environment and sketched out the story of our successful implementation of typo3 at the University of Missouri-Rolla. These comments were eerily prescient in light of what eventually happened to our team at UMR. Despite our CMS's complete success, we earned the enmity of the IT department for succeeding in 5 months where they hadn't in several years and the dislike of our new boss who didn't want any accomplishments of the former Vice Chancellor to remain in place. We were disbanded and the IT department is now going to spend at least $1.2 million on a DMS (Documentum) which cannot even approach our system's functionality and ease of use. All you open-source CMS developers take note--despite the opportunities that exist in the academic arena, choose your situations carefully. The politics will be far more difficult than the technical problems you might encounter. Academia is entrenched in the old world at this point, not the new.
Posted by: Jennifer Lynch | July 25, 2004 10:54 AM
good post, i forwarded it to my friends
Posted by: rent games | November 29, 2004 11:45 AM