Re: SUO: Re: Industry takeover
Adam,
Content is very different from formats and
interfaces. The more content you put into
an ontology, the more likely it will need to
be updated, revised, and extended frequently.
> SUMO is stable at this point, but not
> entirely static. I would envision that we
> could reach a standard, but that like many
> standards, it would need periodic, if minor,
> revision.
As I have mentioned in other notes, stability
is a questionable goal. In IBM jargon, the term
"functionally stabilized" is applied to software
for which maintenance has been terminated -- in
other words, it is a euphemism for "dead end".
If SUMO has been stable, that is most likely
a sign that nobody is using it. If they had
been using it for serious applications, you
would be getting a steady stream of updates,
revisions, extensions, and multitudinous
requests for help in making them.
Furthermore, the whole point of an SUO is
to support much larger subhierarchies of
lower-level categories (which, as the Applied
Semantics people claim, extend into the
millions). If people were actively linking
SUMO to all those hierarchies, you would
have been getting an endless stream of error
reports about conflicts with and among all
the incompatible ontologies that are out there.
These are the reasons why I said that the SUO
must support mechanisms, such as IFF or something
similar, to support revisions, extensions, mergers,
subsetting, etc. Any and every application of SUMO
will require them.
And if you aren't getting such requests on a daily
basis, that is not a good sign.
John