Re: [ALL] proposed agenda 22 July telecon

* Miles, AJ (Alistair)  <A.J.Miles@rl.ac.uk> [2004-07-22 15:19+0100]

> In a nutshell, what I'm hoping for is to be able to launch 'SKOS-Core phase
> 2 development' as soon as possible, with the full backing and involvement of
> this WG, then a furious couple of months of raising issues and trying to
> solve them, culminating in a couple of notes. 

Sounds good to me.

> With regards to the papers from Amsterdam [2] and Maryland [3] on thesauri
> and semweb, I would definitely like to draw on this work and think it is
> extremely valuable, but I'm not sure exactly how to fit it in initially,
> primarily because it deals with 'thesauri' that are not particularly
> 'thesaurus-like' (NCI and MeSH are semi-ontologies, and Wordnet is Wordnet).

Yep. Some account of that, initially brief, in our published docs would
be good. Perhaps in the context of mentioning this WG's Wordnet-related
efforts, distinctions between term-centric and concept-centric models of
thesauri, etc? I just wanted to note that those pieces of work should
definitely be cited and considered...


 
> Perhaps the proposed 'Guide to Using SKOS-Core for Thesauri' note could be
> divided into a 'Quick Start' section and an 'Advanced Features' section ...
> with some parts of the 'Advanced Features' section inspired by the Amsterdam
> and Maryland work?  

Sounds plausible...

> Just as a thought for the longer-term ... with things like Mesh and NCI, we
> get into the hazy world of the relationship between thesauri and ontologies,
> modelling in RDF things that are half-way in between, and also the issue of
> migrating thesauri to ontologies - areas that probably deserve special
> attention (and their own note(s)?)  

Yes re hazy area. I think there is definite scope for us to clarify the 
advantages/disadvantages of both worlds, and how they relate. But that
it'll be much easier for us to do this once we have SKOS or something
quite like it, as well as a Wordnet-as-RDF model. I know my apps will
continue using the Wordnet approach that maps terms and synsets into
classes, maybe there'll be somewhere that design choice could be
explained too.

> But I feel like there are lots of basic problems for us to solve first -
> like a well-documented RDF schema that can cope with all the common features
> of the more standard thesauri.  

Yes. Doing this will make the other issues a lot clearer (eg. we can
write test cases).

cheers,

Dan

Received on Thursday, 22 July 2004 10:30:07 UTC