Re: The horse race

"JOG" <jog_at_cs.nott.ac.uk> wrote:

>Mark Johnson wrote:
>> "JOG" <jog_at_cs.nott.ac.uk> wrote:

>> Codd spoke of tables, as I understand it. I'm not sure a table
>> corresponds to a relation, however.

>Not really, he talked about relationships in the form of relations - a
>table is just a representation. I'm sure you have looked at it, but
>just in case you have not the best source of course is Codd's original
>paper, "A relational model for large shared databanks", 1970,
>Communications of the ACM.

If those 'rules' are not his rules, then so be it. But you should know that 'his rules' read as follows on various sites:

"All information in a relational database is represented explicitly at the logical level and in exactly one way - by values in tables. "

"Each and every datum (atomic value) in a relational database is guaranteed to be logically accessible by resorting to a combination of table name, primary key value, and column name. "

There's those tables, again.

>> >former ordering of elements within the tuple does matter

>> Because they are ordered by position.

>indeed - I tend to refer to planetmath when online for formal
>mathematical definitions - it's a good site.

I was agreeing with you. And then:

>> >and in the latter it does not due to column names

>> Because they are named, which corresponds to a keyword/value pairing.

>indeed each tuple is finite partial map - I would refer you to Jan
>Hidders previous posts as he comes across extremely knowledgeable in
>this area.

"indeed" this is the reason, as I understand it, for Codd saying, above - "and column name." Received on Fri Feb 24 2006 - 01:54:55 CET