If I were to list the top 10 innovations that catapulted the IT industry into a trillion dollar industry, then definitely we would list the mainframe, the microprocessor, the PC, the netscape browser, the myriad PARC innovations, ... I would assert that in the top 5 would be the invention of SQL. Why? Because I think that it allowed for the separation of data and logic, and spurred innovation above and below the layer. While estimates vary, probably $100B of software revenue got generated as result.
The interesting thing about the approx $25B relational database market in the support of SQL really blossomed after the invention of SQL. Hence, different database providers (Informix, Oracle, IBM, Sybase) really tried to differentiate themselves below the layer. As I have said before, if the same SQL is given to Oracle and DB2 on the same data, they better give the same answer, otherwise one (which one? :) or both are not performing correctly. Of course, speeds and feeds, manageability, tieins to programming, tools, availability etc. are fine for differentiation. Now the story is not exactly this ideal, but you get the idea. In other words, databases under SQL embody the user dictum, "Do what I say."
At the other extreme, consider search. The same query on the same set of documents better not produce the same answer in google, yahoo or msn. Otherwise their business models around monetization through better search results go out the window. In other words, search engines embody the user dictum, "Do what I mean." Of course, the interpretation of "mean" is left as an exercise for the implementor :)
Content Management systems -- IBM's CM, Filenet, Documentum, OpenText etc. fall somewhere in between. Much through the history of the content management market, there have been no standards -- so while the concepts have been consistent, the details have been different across the different systems. More interestingly, the definition of what constitutes a content management system has also evolved. Unlike databases, where the main value proposition was to "put" and "get" records (with all the correctness, of course), content management systems have taken on "records management", "archiving", "collaboration", and more recently, "content driven business processes." Thus the boundary between "data" and "logic" has been drawn differently in content management systems. So whatever is the steady state, even if any standards such as JSR-170 or its derivatives become pervasive, the steady state of content management systems will be different than that of database systems.
In some other ways though, content management systems are following a trajectory similar to databases. In databases, first came transaction processing, then business intelligence (i.e. first run your business correctly, then optimize it). Today's sweet spot for content is definitely transaction processing -- e.g. loan origination. But either thorugh compliance, or through issues of optimization, we are beginning to see business intelligence applications and use cases around content. And just like in databases, the structures to make it happen will build upon the transactional structures, but add new capabilities (such as much more integrated search and discovery).
I find the different but somewhat similar trajectories of databases, content and search quite fascinating. Am I the only one?
Nice Blog.
CMS vendors have been very slow in embracing standards. By *not* following standards, I suspect they think they'll be better able to differentiate!
On another note, the number of vendors who call themselves as CMS vendors is just too huge as compared to database vendors and hence its all the more difficult to achieve standardization.
/a
http://apoorv.info
Posted by: apoorv | May 22, 2007 at 10:27 AM