Re: 4 the FAQ: Are Commercial DBMS Truly Relational?

"Marshall Spight" <mspight_at_dnai.com> wrote in message news:dE%9d.363785$Fg5.239889_at_attbi_s53...
> It's just appallingly verbose, don't you think? I compare mathematical
> notation, (which is probably too brief...) with SQL, and it's quite
> enlightening.

No, I actually think it's too terse. Must be different points of view.

> I think a good syntax is a worthwhile endevour in and of itself. I
> wouldn't say it's as important as good semantics, but it's probably
> *almost* as important.

If I had to make a choice, I'd choose semantics too. But I don't see it as an "either/or" kind of thing.

> Every time I get an error from JDBC like "Error in SQL syntax: so such
> column 'CutsomerId' in table 'Customers'" at runtime, I want to retch.
> Such a construct shouldn't compile.

Agreed.

But you don't get that behaviour with precompilers. I think I've already remarked about that to you.
The precompiler validates the SQL syntax, and validates table and column names at precompile time.

Of course, if you generate dynamic SQL at runtime, now you're stuck with error detection at runtime. But if all your SQL is hard coded, you're not stuck. Received on Sun Oct 10 2004 - 05:16:22 CEST