Re: SUO: Procedural question
> [Frank Farance]
>
> The IEEE requirement for consensus building is 75% in sponsor ballot.
> Sometimes, as Jim pointed out, projects discover that they may need to split
> into separate, independent projects. Each project has its own 75% consensus
> requirement. If people want to continue SUMO, OpenCyc, IFF, etc., as separate
> projects, they may. If they want to continue them in the same project,
> they'll have to get to some agreement (merge, face-off, something new, etc.)
> that gets to 75% <-- no different than another standards projects with
> competing ideas/technologies.
Taking Frank's characterization of the end game (the "when" question has not
been addressed) as a good first approximation, we have two possibilities for
arriving at some kind of standard: Merge or Face-off
Merge won't work. If it means union we'll end up with with what I call the
"Flintstone" scenario: Fred and Barney buy a boat together; Fred wants to
call it "Nautical Lady" and Barney "Queen of the Sea", so they merge the
names and call it "Nausea". If it means intersection, we'll end up with
almost nothing except for a one-node taxonomy - Entity or some such.
How would a face-off work? I can't even begin to imagine what the criteria
for such a contest would be, such that the criteria would be universally
agreed upon and seen as fair.
The third alternative mentioned by Frank, "something new", is really not an
option. If something new comes then we just end up with N+1 candidates
where we have N now, and we will either merge them or face them off. Given
that there's emotional as well as technical investment in the current
options I don't think something new could replace the candidates without a
face-off -- so we're back to two choices.
We have also:
> [Jean-Luc Delatre]
>
> ROTFL!
>
> What makes you think that SUMO (or anything like it for that matter...)
> could build consensus?
JL later backed off from what he called "elliptical and hasty wording" but
I'm not sure he should have. His point is an excellent one. Whether in the
broad context of the world or the narrow context of this group, I'm doubtful
consensus can be built because there is simply no basis for building it.
The wild divergence of candidates so far (I mean IFF vs OpenCyc/SUMO) is
evidence that the "spec" for this activity is not at all well-defined. And
if the spec isn't even constrained enough to focus what's being built here,
why should we have any reason to believe that it's constrained enough to
allow us to come to some rational decision on what should win?
> [Erik Larson to Frank Farance]
>
> The point you make here about SUO goals isn't just obvious to me. Perhaps it
> would be beneficial to the group for someone to dig up the relevant "scope and
> goals" sections and explicate exactly what the differences are supposed to be.
> For my part, claiming that SUO work "doesn't set itself to be adopted (large
> scale) by the web" is tantamount to relegating it to obscurity. What, after
> all, is the point of providing ontologies to encourage application
> interoperability if none of this is to be directed at improving the Web? And
> one can only assume that such improvements, were they forthcoming by SUO,
> would be on a "large-scale".
Erik makes an excellent point. SUO can't hope to hide any longer in a
utopian GOFAI world. Whether we like it or not, the W3C is working hard to
develop "ontology" standards that will draw much more attention than
anything this group is doing. Something ought to be done and fast or we
risk being OBE.
I would suggest that what is needed, if this group is to remain relevant at
all, is to do some garbage collection and to rethink the original goals of
SUO. Here's a start:
1) How would SUO mesh with emerging web standards? This is no trivial
question - it goes way beyond syntax.
2) What properties do all SUO entries have to have?
3) What are the concrete testable criteria for judging when we have
something sufficient to fill the role of a standard? In other words,
where's the finish line?
4) We can't do this forever. When is the cutoff date?
5) By what criteria will the entries satisfying points 1, 2, and 3 be
judged when we reach the date set forth by 4?
Anyway, that should be enough to think of for now.
.bill
--
Bill Andersen
Chief Scientist, Ontology Works
1132 Annapolis Road, Suite 104
Odenton, Maryland, 21113
Mobile: 443-858-6444
Office: 410-674-7600
Web: http://www.ontologyworks.com