Re: A new proof of the superiority of set oriented approaches: numerical/time serie linear interpolation

From: paul c <toledobythesea_at_oohay.ac>
Date: Thu, 03 May 2007 13:52:05 GMT
Message-ID: <9Ol_h.156067$DE1.145808@pd7urf2no>

Jon Heggland wrote:
> paul c wrote:
>

>>(I was only trying to counter what I thought was an attempt to make
>>variables inherent in RT as opposed to the language people apply it with.)

>
>
> That may be an admirable goal---the idea of a functional
> programming-style database seems to surface from time to time, though I
> have yet to see anything concrete---but to talk vaguely about names for
> relations (or relation values), that may refer to different relations at
> different times, is (in my opinion) just trying to describe a variable
> without actually using the "v" word.
> ...

Oh sure, the only bogey-man I have about the v-things is when a procedural style forces me to be imaginative about names when it's all I can do to keep the sequence and relationships in my head. Maybe it's sequence that bothers me even more. I would just as soon have the dbms read my program twice and put it order for me.

> As for variables being inherent to RT ... You can discuss relational
> algebra without mentioning variables; the algebra is only concerned with
> values. But as soon as updates (whether INSERTs, UPDATEs, DELETEs or
> assignments) come into the picture, I don't see how you can avoid variables.

(Wasn't suggesting that primes are anything more than notational convenience.)

p Received on Thu May 03 2007 - 08:52:05 CDT