Re: Design Issues (1) - Specialised vocab vs. extensible modular voca bs?

Hi Alistair,

A very generic base vocabularly such as the one you suggest sounds fine. I 
just wouldn't describe it as a common core for KOS languages.

Just look at how much intellecual effort was required just to unify RDFS 
and DAML enough to get the compromise that is OWL - for example, with 
OWL/DL and OWL/Lite not being supersets of RDFS. If the foundation starts 
to express anything richer than labels then the semantics constrains what 
you can cleanly build on top. I'm not sure you can have all of simplicity, 
generality and power. You are opting for the first two, which seems reasonable.

What you have is fine as a core for "things that are like thesauri", which, 
as I understand it, is the aim of the work package. But, for example, I'm 
not sure how well it corresponds to topic maps which, I think, have a 
different way of separating concept from indicator. For the very rich 
representations like conceptual maps then this core covers so little as to 
not really make much difference either way.

Cheers,
Dave

Miles, AJ (Alistair) wrote:

> Hi Dave,
> 
> Yes I was using 'KOS' for everything, including ontologies, thesauri,
> classification schemes, taxonomies, glossaries, dictionaries, topic maps,
> semantic nets, concepts maps etc...
> 
> You make a very good point I think.   
> 
> The design goal I had in mind when I suggested modular vocabs is this: I
> wanted to avoid the worst case scenario, which is that we have one RDF vocab
> for ontologies, one for thesauri, one for classification schemes and so on
> for each type of KOS, and there is no overlap or reuse of components between
> the vocabs.  This is in fact close to what we have right now (e.g. OWL for
> ontologies, TIF for thesauri, LOM-CLS for classification schemes).  
> 
> If the core vocab for all KOS is nothing more than RDF(S) then great.  We
> have OWL already.  So the gap that remains to be filled is for all KOS that
> are not ontologies.  So I imagine soks-core to be aimed at the common
> features of all KOS that are not ontologies, re-using RDF RDFS and OWL as
> much as possible.  Then build an extension for all thesauri and things that
> are like thesauri.  
> 
> To summarise my design goals: maximise re-use, maximise interoperability,
> minimise complexity.  
> 
> To give an idea of how I imagine these modular vocabs fitting together, have
> a look at
> 
> http://www.w3c.rl.ac.uk/SWAD/deliverables/8.1_0_3.html 
> 
> Yours,
> 
> Al.
> 
> -----Original Message-----
> From: Dave Reynolds [mailto:der@hplb.hpl.hp.com]
> Sent: 04 November 2003 10:48
> To: Miles, AJ (Alistair) 
> Cc: 'public-esw-thes@w3.org'
> Subject: Re: Design Issues (1) - Specialised vocab vs. extensible
> modular voca bs?
> 
> 
> 
> Hi,
> 
> I can't access the Wiki at the moment, and in any case don't know how you
> want 
> comments, so I'll stick to email.
> 
> By KOS do you mean "Knowledge Organization Scheme"?
> What does that cover? Are you including ontologies in there?
> 
> If the answer to the last question is "yes" then I suggest avoiding 3. A
> common 
> core for all different thesaurus and ontology schemes is in danger of being
> so 
> generic as to be little more than raw RDF(S).
> 
> Dave
> 
> Miles, AJ (Alistair) wrote:
> 
> 
>>I've added this issue to the discussion on the RDF Thesaurus wiki page
>><http://esw.w3.org/topic/RdfThesaurus>
>>
>>Here is a summary:-
>>
>>
>>Issue 1 - Specialised vocab vs. extensible modular vocabs?
>>
>>
>>Although most thesauri are pretty similar, there are important variations,
>>and many thesauri deviate from the standards. Also, thesauri are very
>>similar to other KOS e.g. classification schemes, taxonomis, topic maps.
> 
> How
> 
>>do we cope with this? 
>>
>>Option 1 - Define a specialised vocabulary that covers only thesauri that
>>comply with the standards. 
>>
>>Option 2 - Define a core vocab that captures what is common to all
> 
> thesauri.
> 
>>Then define extension modules to cope with different flavours of thesauri.
> 
> 
>>Option 3 - Define a core vocab that captures what is common to all KOS
>>(thesauri, taxonomies, classification schemes, topic maps etc.). Define
>>first level extension module for thesauri. Define second level extension
> 
> for
> 
>>flavours. 
>>
>>=== Comments on Issue 1 === 
>>
>>AJM>> 
>>
>>What we did previously ( [WWW]early draft of 8.1
>><http://www.w3c.rl.ac.uk/SWAD/deliv81.htm>) was half way between (1) and
>>(2). 
>>
>>I would like to go for (3), but am prepared to backtrack towards (2),
> 
> which
> 
>>may happen when we hit interop with this and OWL. (3) Would mean we have a
>>way of fitting all these KOS together on the semantic web, which would be
> 
> a
> 
>>good thing. 
>>
>>Going for (3) means we have to define a core vocab. I've kind of assumed
>>this is what we are doing (tell me if you think it's a bad idea), and
> 
> issues
> 
>>below relate first to this core vocab. We need a name for this core vocab,
>>so at least we can refer to it. For now, I'm going to call it the core
>>vocab. In code, I'm using the prefix soks. Why soks? Short for SuperKOS!
> 
> Got
> 
>>any ideas about a better name? 
>>
>>
>>
>>CCLRC - Rutherford Appleton Laboratory
>>Building R1 Room 1.60
>>Fermi Avenue
>>Chilton
>>Didcot
>>Oxfordshire OX11 0QX
>>United Kingdom
>>
>>Email:        a.j.miles@rl.ac.uk
>>Telephone: +44 (0)1235 445440
>>
>>

Received on Thursday, 6 November 2003 09:03:57 UTC