Re: The IDS, the EDS and the DBMS

From: Laconic2 <laconic2_at_comcast.net>
Date: Sat, 4 Sep 2004 08:19:06 -0400
Message-ID: <WJSdnc4gYJp_L6TcRVn-qQ_at_comcast.com>


"Marshall Spight" <mspight_at_dnai.com> wrote in message news:CAa_c.287375$eM2.249312_at_attbi_s51...
> "Tony Andrews" <andrewst_at_onetel.com> wrote in message
news:cha6ah$b9c_at_odak26.prod.google.com...
> > Marshall Spight wrote:
> > > I think it happens about the time they realize the bogosity of ORM.
> >
> > Do you mean "Object Role Modelling" (www.orm.net), or is there another
> > meaning of ORM that I'm missing? If so, I thought that was a
> > relatively little-known movement to supplant E-R analysis tools, not
> > something most application programmers would be interested in.
>
> Uh, sorry for being unclear. I mean "object-relational mapping." You
> know, those million-and-one projects on sourceforge to provide you
> with an object-based interface to your relational data, which comes
> merely at the cost of crippling the kinds of queries you can make.
>
>
> Marshall
>

What I think would be interesting would be a mapping that goes the other way.

Instead of an object-based interface to your relational data, it would provide
SQL type access to object oriented data. The object interface would be natural and easy,
because it would just be an object talking to another object. No big deal.

But the object that served as the interface would "remember" things it had been told, in such a way
that a later structured query "makes sense". I don't know how you do that without imposing some sort of structure on the data at store time. But maybe someone else can figure it out.

It shouldn't be any harder than using wormholes to speed up shipping packages. Received on Sat Sep 04 2004 - 14:19:06 CEST

Original text of this message