Re: Clean Object Class Design -- What is it?

"Bob Badour" <bbadour_at_golden.net> wrote in message news:<Zi%17.4205$9m1.94518425_at_radon.golden.net>...
> >Bob,
> >
> >Maybe you should talk to this guy at Microsoft:
> >
> >http://buffy.eecs.berkeley.edu/Seminars/Old.2001/Apr/010412.hamilton.html
> >
> >He talks about the language execution environment being merged with
> >that of the server. Perhaps not exactly what you are looking for, but
> >maybe a step in the direction.
> >
> >akmal
>
> Hi Akmal,
>
> From the abstract, I cannot tell if the MSFT guy is talking about the same
> thing I am. I suspect he wants to expose DBMS internals to users, which is a
> bad idea. In the end, doing so would just enable MSFT to push their work and
> the work of the DBMS onto users.
>
> The languages supported for type specification do not seem like they would
> lend themselves well to optimization by the DBMS, and I am a little leery
> since all of the mentioned languages rely on references and dynamic memory
> allocation even for very simple object types.
>
> By any chance, did you attend the seminar?
>
> I have not investigated .NET or CLR much. It would be nice if MSFT built
> higher-level abstractions, like relations, into CLR, but I doubt anyone
> there has considered the idea.
>
> In short, I would be happier if he were talking about merging relational
> technology into the language execution environment than vice versa.
>
> Cheers,
> Bob

Bob,

Unfortunately, I didn't attend the seminar as I'm based in London. However, my colleague, Paul Brown at IBM Informix Labs did attend. I will ask him about his impressions in detail. Paul Brown you may have heard of, as he worked with Mike Stonebraker on Postgres at UCB. Paul is interested in database extensibility and also considers many of the theoretical and practical issues. I understand that IBM have also been working on various initiatives over the past few years. Oracle too, although someone that I admired greatly, David Beech, retired earlier this year.

I confess, I haven't come across anything that discusses merging relational with the language environment. This could be quite interesting. Maybe I'll try a search at Michael Ley's web site. Or if someone knows of some research ... :)

akmal Received on Mon Jul 09 2001 - 00:59:05 CEST