My take on the discussion yesterday is the following. I agree
with Steve Ray that it could be useful to achieve a more 'commonsense'
description of ontologies that would enable them to be sold
more readily to those without the technical backup. I am
sure many of us have used variants of the very useful
overhead from LOA and the Wonderweb project showing the
dimension ranging from indexes and glossaries at one
end to full formalised axiomatised ontologies at the
other. The update of this in Michael Uschold's overhead
on the wiki is a further example: (01)
http://ontolog.cim3.net/cgi-bin/wiki.pl?MichaelUschold (02)
These establish a clearly understandable (or apparently so)
characterisation that people can situate their interests on.
I presume that a characterisation such as Steve was suggesting
would move us further in the direction of having a finer
articulation of the positions on this and other similar
dimensions. (03)
I would not agree, however, that such an endeavor should
be undertaken *before* looking at more formal characterisations.
We can already say a lot more about the dimensions taking
us from index to Ontology adopting what is known from the
formalisation side. But, conversely, the fact that we do
not have the final word on a formalised characterisation of
the space of ontologies and the simple dimension is still
quite useful, demonstrates that we need not wait for the
final formal word before having useful taxonomies (folksonomies? :-)
of ontologies themselves. Therefore I see a need for
(at least) three parallel working group efforts, that
will each appeal to different (but hopefully overlapping)
participants. These could be pursued in parallel with
as frequent as possible reporting back to see how soon
they can start being combined. (04)
(1) Characterisation of 'kinds of ontology' building on
whatever dimensions of characterisation appear useful.
These could start with the index-glossary-taxonomy-
Ontology continuum and add as seems useful. (05)
(2) Characterisation of kinds of ontology building on
the formal structuring mechanisms employed (one module,
many modules; modules of the same expressivity,
modules of different expressivities; ...), on the
formal languages employed in each of these
modules, on ways of characterising
the concrete use made of the expressive resources
available (e.g., proving that some ontology X actually
lies within some restricted modal logic rather than
the full first order logic casually claimed by its
developers; or that a fragment lies within some other
sublanguage, etc.), and then perhaps notions of size. (06)
(3) Characterisation of use cases, working through a classification
of what kinds of purposes, functionalities, users,
are involved. (07)
I would expect that there are going to be many points
of contact and, finally, even semiformaliseable connections/relations
between the three activities. But since there is much to
do in each of these areas before it is quite clear what is
going to be talked about even internally, mixing
the three from the start may
be a way of having a lot of very crossed
purposes/wires. (08)
This might be one way of leading up to a very nice
'deliverable' by the time of the April summit. (09)
Thoughts? (010)
best,
John B. (011)
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2007/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2007
Community Portal: http://ontolog.cim3.net/ (012)
|