Re: a union is always a join!

From: paul c <>
Date: Fri, 30 Jan 2009 20:18:04 GMT
Message-ID: <0gJgl.8119$PH1.4873_at_edtnps82> wrote:
> On Jan 8, 11:21 am, (rpost) wrote:
>>  I would
>> venture that the first thing any implementer will attempt to do
>> is rewrite all of Darwen's expressions back to relational algebra as
>> far as possible, and throw errors on the remainder.

> You use "venture", "attempt" and "as far as possible" but the
> semantics are clear.
> ...

Perhaps a neophyte would try that. If he did it a second time though, one would have to question his aptitude for such work, for that matter, any kind of mental work. Apart from predicting and validating results, the chief use of an underlying algebra to define a language is as a means to prove logical optimizations. (I have not seen any dbms that has the kind of exact and formal definitions as Tutorial D has, but even that could go further when it comes to constraints. Most dbms's seem to rely on informal definitions, if any.) Received on Fri Jan 30 2009 - 14:18:04 CST

Original text of this message