Re: what are keys and surrogates?

Marshall wrote:

>On Jan 4, 1:37 pm, rp..._at_pcwin518.campus.tue.nl (rpost) wrote:
>> JOG wrote:
>> [buncha stuff]
>
>I'm jumping in rather late in the discussion, but I want to say
>that I'm not at all sure that the two of you are using words in
>the same way. In particular, I don't think "hidden" means the
>same thing when Reinier uses it as when Jim uses it. (And I
>think there may be other problematic terms as well.)

I just posted a reply with the same conclusion.

>In general, I find Reinier's terminology to be unfamiliar,
>although the concepts he's describing aren't.

OK, then I'll just need to read up on terminology. I'll see if I can get a hold of that book by Date.

>I would recommend separating out the idea of a datatype
>that supports only construction and equality from other
>ideas such as keys. I would also recommend being
>quite explicit about definitions.

"My" definitions are here:

  http://groups.google.com/group/comp.databases.theory/msg/99f260782ff7b380

-- 
Reinier