Re: Generic Modeling

From: D Guntermann <guntermann_at_hotmail.com>
Date: Sun, 13 Jan 2002 01:17:47 GMT
Message-ID: <GpurLM.I3_at_news.boeing.com>


Hmmm. Could not find the reference in Date's 7th Edition Introduction, though I found bibliographical information concerning proposed or prototype object-oriented data access languages such as OSQL and GEM in the reference section of chapter 24, 'Object Databases'.

I did find information about ConQuer at http://www.orm.net/pdf/ER96.pdf. Many aspects of ConQuer seem appealing in terms of bridging perceived gaps between a conceptual query to either an underlying relational or object-oriented database query (presumably dependent on base structures), independent of the syntax or construction of the underlying query.

Yet, some of the justification of the benefit of ConQuer seems flawed because it is based on a predicate that a "query", especially in the relational sense, is tied to a physicial implementation (e..g. an expansion of a multi-valued attribute to a joined relation, etc.). Unfortunately, in the real world, this is probably true. The level of indirection between the "community level" and the "user level" in terms of independence is usually non-existent, though I would suggest this state of affairs is because of ignorance on the part of database designers/implementors/vendors rather than flaws in relational or object-relational capabilities or services (except for maybe the continued weaknesses in fully addressing view updating).

Is this what Xdb uses? Though it might be possible to map a level of indirection between the 'conceptual' and an underlying hierarchical model, if an underlying hierarchical/network model is constrained to the access paths that have been hard-coded that reflects only a portion of those relationships in terms of the conceptual query, then it seems that it would be difficult in terms of performance, or even functionality, for an object-oriented data access language to have the flexibililty to materialize the conceptual queries being presented at the user level.

Dan

<MSherrill_at_compuserve.com> wrote in message news:3c3eff76.2744055_at_news.compuserve.com...
> On Sat, 5 Jan 2002 19:43:04 -0800, "Daniel Guntermann"
> <guntermann_at_uswest.net> wrote:
>
> >Is there a data access language or syntax that would allow a somewhat
rapid
> >construction of this query?
>
> I think ConQuer is supposed to do this. ConQuer was developed by
> Terry Halpin (of ORM fame), now an employee of Microsoft. I don't
> think it's been released, but I could be wrong. IIRC, there are
> references to this in Date's _Introduction ..._, 7th ed.
>
> --
> Mike Sherrill
> Information Management Systems
Received on Sun Jan 13 2002 - 02:17:47 CET

Original text of this message