Re: In an RDBMS, what does "Data" mean?

Dawn M. Wolthuis wrote:
>
> Once again thanks for the link, Jan! I have the feeling that you know
> everything I wish I knew. I heard of the Functional Data Model only after I
> looked at the relational model then looked at PICK and tried to figure out
> what PICK was. Every PICK file is a mathematical function, including the
> Vocabulary file for the entire system.

So, a PICK database is kinda like a relational database restricted to binary relations and a mandatory key constraint on the first column then? :-P

> So, I googled for "functional data
> model" and was surprised to get quite a few hits. I bookmarked a bunch of
> them, but have not read much. I doubt that PICK fits whatever the FDM is,
> but since I'm referring to it as a Functional Database (compared to a
> Relational Database), I definitely will need to figure out just what the FDM
> is.

Actually, the original is quite readable.

D.W. Shipman. The Functional Data Model and the Data Language DAPLEX. ACM Transactions on Database Systems, 6(1):140-173, 1981.

But I wouldn't say that there is something really profoundly new about it. It's sort of a version of the Entity-Relationship Model but with no distinction between Entities and Relationships (I consider that a feature) and more emphasis on the roles (i.e. the relationship between Entities and Relationships). But it's simple, elegant, well-formalized, effective and is untuitively understandable in a natural way. Very nice for conceptual modeling, for example.

> I read the link above, but I must be in one of those cloudy states because I
> didn't end up with a much better idea of what the FDM was.

Yeah, sorry about that. Just let me know if you cannot get your hands on the Shipman paper. Perhaps I can arrange something.

> I have this link
> on my "to read" list http://coronet.iicm.edu/dm/scripts/lesson06.htm along
> with some others, so eventually I'll get to it.

Hmm, it looks a bit odd. They talk about multi-valued functions returning lists, and I believe that's not quite correct, it should be sets.

Did I already mention that there are new exciting papers on normalization theory in list-based data models (as opposed to set-based data models)? Pretty profound stuff.

• Jan Hidders