Dear Mr. Oram,
I read your article with interest and would like to share some thoughts in response.
I agree with Mr.Bosworth that putting more and more features into a database engine is the wrong approach. In order to describe why, I have to take a step back and look at the relation between data, information and knowledge.
Data is managed by DBMS, information by knowledge bases, faq systems and the likes, knowledge - well, there are a few out there which actually correlate information to allow for new conclusions.
To build information- or even knowledge mgmt. functionality in a database system, is to build a system to do something it was not designed for.
Wouldn't it be better to define flexible interfaces and functions (services?) to be able to be used by info/knowl. mgmt. systems ?
For my own experimentation, I am looking at tikiwiki as a system to evaluate knowledge mgmt. with. it is build using mysql (or other dbs) as an underlying dbengine and does a great job.
Knowledge is distributed and unstructured and changing all the time. To discover knowledge, a system would need to be able to be distributed, manage unstructed information and adapt to changes.
Could this be a feature for DBMS ?
CKO InsightKnowledge Inc.