Re: OT: SQL

On May 4, 7:25 am, Charlie-Boo <shymath...@xxxxxxxxx> wrote:
[...]

Well, you say a bunch of stuff but you don't respond to what I said
at all. (In fact it's not clear you even read it.) You're using words
like "file" in different ways at different times, and not clarifying
what you mean by it, despite requests and despite the fact that
that is not a data management term. (Nor a logic one.) You
propose a plethora of ideas. Some of them are sound but not
a good idea: using predicate calculus directly as the query
*language* might possibly make sense if you want a database
that only logicians can use; it buys you nothing expressively.
(SQL is something between an algebra and a calculus; I agree
most strongly with those authors who argue that we need
something that is more algebraic.)

Some of your ideas are just wrong: "Codd et. al. never
figured out how to handle [relations that] have multiple levels
in a hierarchy." Codd proved (in the mathematical sense) that
such are unnecessary, and advocated against nested relations
as a design principle (not a mathematical one); and "et. al"
have written about a bazillion papers on the nested relational
algebra, and even Codd's closest associates today consider
nested relations allowable.

Still other of your ideas are good ones, like materialized
views, that have been incorporated into existing dbms
products over the past couple of decades.

The idea of using preexisting files of arbitrary structure
as part of a dbms is so spectacularly bad I am speechless.

Going on and on about Codd seems pointless to me. Codd
wrote some papers nearly forty years ago that today are
seen as very influential, seminal even, but of course they
leave out the world of things figured out since then. It's
as if I was talking to someone about set theory but they
had only read Cantor's original work.


Marshall
.

Relevant Pages

  • Re: A different definition of MINUS, Part 3
    ... I think around 1972 Codd wrote a proof that the algebra was logically ... equivalent to FOPC, later others corrected a few ... algebra would be a sufficient tool to clarify all aspects or RL. ...
    (comp.databases.theory)
  • Re: A different definition of MINUS, Part 3
    ... DandD work seem to ignore without consequence or awareness of the ... I think around 1972 Codd wrote a proof that the algebra was logically ... Codd never mentionned that FOPC would be sufficient ...
    (comp.databases.theory)
  • Re: A different definition of MINUS, Part 3
    ... I think around 1972 Codd wrote a proof that the algebra was logically ... I think Codd meant that his relational model obeyed both of FOPC and his algebra. ... The personal choice is whether one uses algebra or fopc not only to understand his model but to define it. ...
    (comp.databases.theory)
  • Re: RM and abstract syntax trees
    ... agree that it's not much fun to manipulate them, I wish Codd had said more about nested relations as I have a feeling he spent some time considering them.) ... Sorry I don't have much mastery of conventional syntax, what I mean here is something like R: where is a set of attribute name, attribute type pairs and typeof is swiped from C-language: ...
    (comp.databases.theory)