Re: Do Data Models Need to built on a Mathematical Concept?

In short. I agree with you Costin.

Re-reading my words, I said "much more respectable [ideas may come out of this effort] than you might think". Following Fabian Pascal, you would think no useful ideas could ever arrive, indeed after the experience of the OO extensions to SQL (DB2 in my case), no useful ideas at all showed up (which annoyed me, I thought I might get half-way useable User Defined Types at the very least). This time I am just a little more hopeful, that is all. Maybe it will be 'higher-order' languages (like SchemaSQL). That is certainly one thing that the FP guys have (or are on the way to getting with 'meta-programming' research) and it's one of the things that I want in my relational languages.

BTW I do believe that the _underlying_ goal is laudable. Namely non-relational, non-scalar access to relational data. (albeit on relational terms). Indeed the manifesto suggests ARRAY types that can be associated with relation types. It's just a shame that the promoters of this XML + SQL stuff don't see things in such terms...

Regards
Paul Vernon
Business Intelligence, IBM Global Services Received on Mon Apr 28 2003 - 01:02:49 CEST