Re: table types

"Marshall Spight" <mspight_at_dnai.com> wrote in message news:<6uQLc.141450$%_6.57745_at_attbi_s01>...
> "Leandro Guimaraens Faria Corsetti Dutra" <leandro_at_dutra.fastmail.fm> wrote in message
> news:pan.2004.07.22.12.20.13.373768_at_dutra.fastmail.fm...
> > Em Wed, 21 Jul 2004 16:53:04 +0000, Marshall Spight escreveu:
> >
> > > From the benchmarks I've seen over the years, it seems
> > > like the two most "generally fast" dbms are Oracle and
> > > MySQL
> >
> > Your benchmarks are seriously defective.
>
> I was speaking of dbmss that actually exist, and specific
> implementation techniques.
>
> For your statement to be meaningful, you have to show that
> there exist noticably faster alternatives. Note that saying that
> there might one day be faster alternatives is not relevant to my
> statement, because I am discussing current implementations.
> It is quite difficult to run benchmarks against software that
> hasn't been written yet.

Yes, it can be frustrating round here sometimes, can't it? Perhaps your question belonged in comp.databases rather than comp.databases.theory, but if so the appropriate response is to say so. Too often I see references to "real world" DBMSs like Oracle dismissed with comments like "poor choice of DBMS", when what the poster means is that you should use an imaginary "true RDBMS" instead!  Now if a true RDBMS actually existed, that might be a valid point. But it doesn't (yet). (Yes, I've heard of Alphora!)

I am passionately in favour of true RDBMSs, and am well aware of the reasons for the inferiority of SQL in general, and Oracle et al in particular. But we can't all sit on our arses and wait for Nirvana, we have to deal with what we can get. So we shouldn't be treated as fools for doing so. Received on Fri Jul 23 2004 - 11:39:01 CEST