RE: [VM] Scoping Draft with questions to TF members $swbpd

-----Original Message-----
From: Thomas Baker [mailto:thomas.baker@bi.fhg.de] 
Sent: Monday, September 06, 2004 2:26 AM
To: Uschold, Michael F
Cc: Thomas Baker; SW Best Practices
Subject: Re: [VM] Scoping Draft with questions to TF members

On Fri, Sep 03, 2004 at 12:13:52PM -0700, Uschold, Michael F wrote:
>     Were you thinking one might find a different example for each
>     principle, or find _one_ example to provide a common point of
>     reference to drive _all_ of the principles?
> [MFU] One example to illustrate all the principles is better than lots
> of different ones. Though sometimes it is hard to find a single
example
> to illustate all points.
> [MFU] 

I am assuming you mean a "real" example here - not a concocted
one?  So Dublin Core could in fact be an example, since it
illustrates alot of the points we will want to make?
[MFU] A REAL example can be better, but not always. It can be hard to
find one that so readily illustrates the required points.   See the
notes from the [OEP] task force for good examples of examples [as it
were...].

>     I see the point, though this seems like one of those
>     issues just beyond the scope of the core principles --
>     a good candidate for brief discussion in Section 4, 
>     with pointers to further reading.
> 
> [MFU] Hmmm: I'm thinking that this is a MAJOR POINT that should be
> driven home in a compelling way.  One way to proceed is in effect to
use
> OWL as a way to represent a thesaurus lang: if you do it this way,
these
> are the consequences. If on the other hand, you use OWL in a different
> way, then here are the consequences of THAT decision.  Perhaps I'm
> missing the main purpose of this document? It seems to me that if you
> want to include thesauri, taxonomies, controlled vocabularies etc in a
> Semantic Web-friendly way, then this is about the first question to
ask!
> [MFU]

As I understand it, your point is that OWL should be used
"out of the box" to represent a thesaurus language directly --
rather than using OWL first to represent some ad-hoc language
of thesaurus relations and then, in turn, using that ad-hoc
language to represent the thesaurus.


[MFU] NO NO!  I'm remaining agnoistic. The matter needs looking into.
There may be benefits either way. Or there may be clear preferred
choice. 

This gets back to Danny's point [1]:

> [[
> The goal of this Task Force is to describe best practice for declaring

> and managing terms and term sets (vocabularies) for use in a Semantic 
> Web environment.
> ]]
> 
> One of the first questions a vocabulary author is likely to ask when 
> considering how to declare a term is "rdfs:Class or owl:Class", or to 
> put it another way, "shall I use OWL Full (RDF+S), DL or Lite?"
>  
> I was wondering whether this question will be in scope for [VM], and
if 
> not, how on earth will you avoid it?
> 
> I realise 'cleaning up the mess' has been put on hold as far as [WRLD]

> is concerned [1], but there is no need for anyone to commit to advice 
> pointing towards any (sub-)language exclusively. However there should
be 
> information available on the implications of any choice based on known

> facts. The mess won't clean itself up, quite the opposite if people
are 
> implementing without at least some knowledge of the relative 
> strengths/merits of the alternatives.

It would seem that you and Danny are suggesting we go beyond
saying just:

    "Declare the terms in schemas."

and even beyond:
    
    "Declare the terms in RDFS, OWL Full (RDF+S), DL or Lite."

to say, in addition, something along the lines of:

    "You can declare the terms in RDFS, OWL Full (RDF+S), DL
    or Lite, bearing in mind the following considerations...:"


[MFU] perhaps, but this might be too much work, considering all these
options.  Probably is best to start with ONE good way to proceed. While
doing that, issues and choices will emerge that suggest pros/cons of
different options. Proceed from there.

AFAIC, the further we can go down this road in terms of
consensus, the better.  However, I was assuming we would
_not_ be able to recommend just one schema flavor (because
"good practice" is itself not yet sufficiently clear on this
point), nor would we want to spell out the various alternatives
in-line beyond a certain level of detail.

What do others think?

Tom

[1] http://lists.w3.org/Archives/Public/public-swbp-wg/2004Jun/0064.html

-- 
Dr. Thomas Baker                        Thomas.Baker@izb.fraunhofer.de
Institutszentrum Schloss Birlinghoven         mobile +49-160-9664-2129
Fraunhofer-Gesellschaft                          work +49-30-8109-9027
53754 Sankt Augustin, Germany                    fax +49-2241-144-2352
Personal email: thbaker79@alumni.amherst.edu

Received on Tuesday, 7 September 2004 19:45:44 UTC