Re: Is it possible to build a purely relational database on top of SQL?

From: Christopher Browne <cbbrowne_at_acm.org>
Date: 5 Jul 2004 11:20:38 GMT
Message-ID: <2ksrs6F5ogr4U1_at_uni-berlin.de>


Centuries ago, Nostradamus foresaw when rafe_at_cs.mu.oz.au (Ralph Becket) would write:
> Please forgive me if this question is naive/asked every other
> week (to date I've been unable to find an answer on Google.)
>
> My question: is there any reason why a purely relational database
> interface, in the "Third Manifesto" sense of, say, Tutorial D,
> could not be built on top of any existing commercial SQL DBMS?
> That is, could one build a true RDBMS without having to rewrite
> everything from scratch? (Note, I do not require that such
> an interface be expected to work for an arbitrary DB containing
> unpleasantness such as NULLs and duplicate rows etc.)
>
> If this is possible, has it been done? Are there performance
> issues with such an approach?

It has been done, with Alphora's _Dataphor_ product, which apparently supports Oracle, Microsoft SQL Server, IBM DB/2, and other database systems as 'storage devices.'

You'd have to contact the vendor to determine what, if any, performance issues fall out of that.

-- 
select 'cbbrowne' || '_at_' || 'ntlug.org';
http://cbbrowne.com/info/oses.html
QT adds to  a Linux distribution a level  of licencing complexity that
nullifies one of the major  virtues of Linux: no licencing complexity.
-- <jedi_at_dementia.mishnet>
Received on Mon Jul 05 2004 - 13:20:38 CEST

Original text of this message