Re: a union is always a join!

From: paul c <>
Date: Tue, 24 Mar 2009 01:36:04 GMT
Message-ID: <8OWxl.19369$PH1.18138_at_edtnps82> wrote:
> On Mar 23, 11:09 am, (rpost) wrote:

>>>> Yes, but what I meant to say is that in general, the tuples
>>>> don't really express facts regarding those domain values,
>>>> they just help express information about other domain values.
>>> Tuples do not express facts about domain values. They contain domain
>>> values.
>> Duh.
>> True, I didn't express myself clearly enough.
>> But I don't need your lecturing.

> Of course I know you know tuples contain domain values.
> What you didn't seem to know is that they do not express
> information about domain values, since you wrote the
> opposite. I'm only being precise and basic to justify my
> points clearly. I guess you think I'm too basic. But I think
> that many things you write contradict basics, and that thus
> basics are relevant to my reply.
>> My point is that in a relational
>> the tuples of a relation often correspond not just to the
>> propositions of an associated predicate, but to observations; explicitly
>> asserted, rather than derived information.  To propositional logic,
>> it's all the same, of course.

> Along the way you have said (along with a lot of other stuff
> I contradict) that there is a distinction relevant to the user
> between relations that observe changing things, those that
> observe unchanging things and those that derive from
> these; and that it is relevant to the user how any of these
> are implemented. And I have said that there isn't. Other
> than what the observing (changing and constant) relations
> represent, to the *user* it's all the same. And treating them
> them the same eases programming.
> philip

I just want to comment on one point, users can think whatever they want (when they away from the desk as it were), but when they are operating the db, the only view they should take is that of the intentions of the data design. If the design doesn't concern itself with recording changes (as opposed to the result of recording changes), eg., if the design doesn't handle 'before' and 'after', then a user who concerns himself with that is a mystic!

Obviously a data design is just a mechanical reflection of one or more chosen abstractions of reality. All mystics should make a vow to remember daily what McCarthy said about submarines trying to swim.

Mystics are like Chickenman, they turn up all the time, no matter where you go. It is a shame that they have started to breed with technocrats, now neither label has any meaning. Received on Tue Mar 24 2009 - 02:36:04 CET

Original text of this message