Re: Multiplicity, Change and MV

JOG wrote:
<snip>
> In the area I work in, changes in cardinality and table schema are
> rife. Scientific data, especially the way we want to contextualize and
> record it, is changing constantly (this is one of the reasons that the
> hideous EAV model got so much funding). Perhaps I just have to accept
> structural and hence query change via the RM, but perhaps there is a
> solution that will suit my theoretic demands as well as being
> practically effective?

I think that is close to the question I've been asking. I saw products that worked that had little theory to back them up and products backed by a wealth of papers (and theory) that are a PITA to work with. I agree on the EAV, by the way.

I would guess that using products I have little experience with such as Berkeley-DB and Cache' or products I have more experience with such as plain-old indexed sequential files (e.g. VSAM), Pick, and maybe even IMS could all typically yield more effective, efficient, and less expensive solutions for the long haul than the average SQL-DBMS solution. There is some good theory backing non-first normal form, but it doesn't match up well with these solutions from what I have seen so far.

I think we need to do more to get our theory from our practice rather than the other way around. We have an appalling lack of emperical data in our profession.

I'm not sure I fully understand your question, but if I did, then I like the question. Cheers! --dawn Received on Tue Apr 11 2006 - 04:16:21 CEST