Re: Is relational theory irrelevant? (was Re: Dreaming About Redesigning SQL)

"Paul Vernon" <paul.vernon_at_ukk.ibmm.comm> wrote in message news:bpb52a$hde$1_at_gazette.almaden.ibm.com...
> "Mikito Harakiri" <mikharakiri_at_yahoo.com> wrote in message
> > The reference, please. I find it hard to believe that differences in
> > NULL treatment and Multiset Semantics can result in considerable gap
> > in expressibility betwen Relational Algebra/D4 and SQL. (What else can
> > be the source of D4 superiority?).
>
> Hugh Darwen opinions that dependence on column ordering is SQLs third most
> severe mistake (after the two you mention). So that's another source of
> superioity. See here for more:
>
>
http://www.hughdarwen.freeola.com/TheThirdManifesto.web/Importance-of-Column -Names.pdf
>

Is the point of this article is that SQL got some of its features not in the initial design, but as a result of evolution? It doesn't really matter now, because column expressions could be named in SQL, and in some cases DBMS would force user to do so. Therefore, there is no really any gap in expressivity there, and Hugh's appeal to "relational incompleteness" is a strawman. Numbers in "order by" clause are disgusting, of course (for example, what if we add them:-), but that's a minor snag. "Having" is redundant but we are not discussing orthogonality of the language here. View declaration syntax is annoying too, but again it doesn't affect expressivity.

I personally like SQL query syntax better, because it's well structured. Predicates are in "where" clause, the relations in the "from", and column expressions in the "select" list. Any of these 3 parts can contain a subquery. Inner view inside "from" block, and/or scalar subquery inside either "select" or "where" clause. Nothing short of elegance. BTW, I recently learned that subquery in the select list was actually borrowed from ... OQL(!). Hug's opinon about (scalar) subqueries being redundant in the "where" clause is the biggest nonsence on his list SQL redundancies. Received on Mon Nov 17 2003 - 22:08:21 CET