Re: Is Ingres yet relational?

Leandro Guimaraens Faria Corsetti Dutra <leandro_at_dutra.fastmail.fm> wrote in message news:<pan.2004.05.26.15.20.48.49473_at_dutra.fastmail.fm>...
> Em Wed, 26 May 2004 07:06:27 -0700, Paul Mason escreveu:
>
> >> I wonder if such a QUEL-only applications would avoid
> >>duplicates, NULLs and the such.
> >
> > If you specifically wrote your application that way it would. NULLs and
> > duplicates exist but you're not forced to use them.
>
> Indeed, but their existence, and that of SQL access,
> inherently hinders optimisation.

I disagree. It makes optimization more difficult but you can use different algorithms when you know that nulls and duplicates are not allowed.

> Actually what I was thinking was the possibility of creating a
> 'relational Ingres' by striping it of SQL, perhaps replacing it by
> Alphora's RealSQL. But perhaps it would be just as easy to backport
> QUEL to PostgreSQL...

It would be a good possiblity if you know very well the source code of Ingres. If not, probably it would be easier to start from scratch, specially if you want to have decent physical independence.

Regards
  Alfredo Received on Thu May 27 2004 - 17:12:19 CEST