W3C home > Mailing lists > Public > www-webont-wg@w3.org > February 2002

Re: REQDOC: Change List from Editors

From: Jeff Heflin <heflin@cse.lehigh.edu>
Date: Wed, 20 Feb 2002 10:58:32 -0500
Message-ID: <3C73C7A8.3A455757@cse.lehigh.edu>
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
CC: www-webont-wg@w3.org
Peter,

Please see my responses to you concerns below:

"Peter F. Patel-Schneider" wrote:
> 
> Hi:
> 
> Here are some more concerns I have with the requirements document.
> 
> 1/ I appears that there will be considerable editing on the document.  I
>    need another pass on the document to see if any new problems have
>    surfaced.

We are working hard to incorporate the changes in the document by the
end of the day. If at the time of the telecon, you feel that you have
not had sufficient time to read the document or that you concerns which
have not been addressed, please raise the issue there.

> 2/ The introduction to the document reads as if the requirements therein
>    are more than cast-in-stone.  It needs to be toned down to note that it
>    is a document that has been produced by fallable humans, who have to be
>    able to fix their mistakes without causing the end of the world as we
>    know it.

Would a statement to the effect that "the Working Group reserves the
right to remove or add requirements" help?

> 3/ I don't understand what
> 
>         Note that it is possible for a revision to change the intended
>         meaning of a term without changing any axioms.
> 
>     is supposed to mean.  Is OWL supposed to know about the intended
>     meaning of terms in ontologies?  If so, how?

No. The statement was meant to say that only humans can know the
intended meaning of terms. Therefore, only humans can truly determine
semantic backward-compatibility of terms, and thus the language needs an
explicit feature to express this. I'll try to make this clearer in the
document.

> 4/ Many of the objectives are not adequately defined.
> 
>    a) Chained properties
>       Where and when are chained properties allowed?
>    b) Variables
>       What do variables mean?  Where can they occur?  What power should
>       they provide?
>    c) Arithmetic primitives
>       What sort of primitives?  Where can they occur?
>    d) String manipulation
>       See above.
>    e) Pre- and post- conditions
>       What do these mean?  Do they require that OWL incorporate a theory of
>       time and action?

I'm not sure if we can provide the kind of details you ask for at this
time. Where these things can occur depends on the structure of the
language that we end up with. A lot of these would be more appropriate
for a rule-based language than a description logic. Of course, if anyone
would like to provide a more detailed description of one or more of
these, please be my guest.


> 5/ Some of the objectives do not fit within a language definition or fall
>    outside of the ontology level
> 
>    a) Integration of digital signatures
>       This is either part of the transport mechanism, or part of the trust
>       level of the semantic web
>    b) Bit-efficient encodings
>       XML is the transport mechanism.  However, nothing prevents agents
>       from implementing a compressed transport mechanism.

I wasn't a big fan of including these myself. However, I kept them as
placeholders because they were important to some members of the group.
Unless these member speak up now, I'll strike them from the document.

ACTION: Integration of digital signatures and bit-efficient encodings
will be removed from the document.

> 6/ Some of the concerns in my previous message appear to not have made it
>    into the list of changes.

Please let me know which concerns these were. One of the points of the
change list is so that you can double-check that we haven't missed
something.

> Peter F. Patel-Schneider
> Bell Labs Research
>
Received on Wednesday, 20 February 2002 10:58:35 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:47 GMT