Re: What are we asking of a data model?

Joe \"Nuke Me Xemu\" Foster wrote:

> "SPeacock" <Xpeacock_at_pacbell.net> wrote in message <news:4001BE3B.3BFF4B97_at_pacbell.net>...
>
> > all of these by using only a single (the same) tuple.
>
> While, of course, the Pure-OO solution gets to embed absolutely
> anything whatsoever inside the "equivalent" objects, including
> arrays, collections, trees, reflection, etc. etc. etc., right?
>
> --
> Joe Foster <mailto:jlfoster%40znet.com> DC8s in Spaace: <http://www.xenu.net/>

Well the best solution is to include them 'logically' within the right relations, but physically they maybe stored someplace that is more appropriate. Again we don't want to mess around too much with the 'physical' stuff in the data model I mentioned. But yeps, the data model has to account for everything. In the OO world too. The model I am talking about is definitely not an OO (just a variation of relational if you look at the UML implementations) nor a relational schema.

And you need a 'data language' to be able to describe those relations...not like a natural language but it has to have characteristics that allow associations just like parts of speech... but a language for data that would support similar characteristics such as the underlying rules of Math or Music, etc. These would be rules supporting the formation of a data model. And not implantation oriented. The datamodel would be able to model the relational/OO and transpose to it. The relational/OO models would not be able to model it. It could in some cases be implemented in relational/OO. But for complex relationships, relational (and sister OO) would be hard pressed to implement the entire model. Received on Mon Jan 12 2004 - 01:24:29 CET