From: Adam Pease ([email protected])
Date: 04/10/02

   I tried to allude to the issue you raised, but maybe did not to justice 
to in the text:

"It should also be easy to see that there is similar utility in just
about every domain: in finance where one might query about all bank
accounts associated with a particular person (whether they are directly
owned by, or held in trust for etc), in logistics where one wants to ask
the rates for shipping to any eastern European city (where no such category
has been predefined and only the countries in eastern Europe are
listed).  "

One problem is that we could cite any number of examples in different 
domains.  Which ones do we choose?  Maybe one way to handle this is to post 
the document as is and develop a suite of illustrative examples that would 
be motivating for different communities.  Would that be a reasonable way to 
proceed?  Would you like to develop a fuller logistics example the document 
could reference?


At 12:51 PM 4/10/2002 -0700, Gio Wiederhold wrote:
>    Perhaps you did not get my comment?
>I feel quite strongly about the first point.
>Date: Mon, 8 Apr 2002 15:37:04 PDT
>From: Gio Wiederhold <[email protected]>
>To: tim finin <[email protected]>
>Cc: Adam Pease <[email protected]>, [email protected]
>In-Reply-To: Your message of Mon, 08 Apr 2002 15:08:16 -0400
>Fcc: ./DAML/reports.m00
>Message-ID: <[email protected]>
>Adam's example is great start.
>To convince DARPA's customers it would be nice to give an example from
>logistics, for instance one  where similar part assemblies are needed
>in a system?
>In general, the motivation is to make E-commerce reliable, and move
>  from browse, read, and paste to automation.
>Can the propery attribute help there?
>Btw., limiting expressive power is not a hindrance, if it doesn't
>limit the customer.  Databases were successful largely because the
>relational model was simple and clear to its customers.
>When research databases added recursion it excited the academics, but
>did nothing much for the customers.  I recall the VLDB paper that
>started "Since no-one can be his own father .. ". I sent the author
>a tape of the country jingle `He's his own Grampa .. which goes on
>telling how the son married the widder of the father, etc.'. One
>counterexample can kill any theorem ... .
> >>>>>>>>>>
>Making examples that require extrapolation from what looks like
>a toy problem to the problems that our intended audience and funders are
>facing makes us look as we are playing in a private sandbox.
>/Gio Wiederhold/

Adam Pease
(650) 424-0500 x571

This archive was generated by hypermail 2.1.4 : 03/26/03 EST