Re: SUO: Procedure of development
Tim,
I think that Frank has given a more extensive response that accords with
standards practice, but I wanted to clarify informally my
intention. Certainly it is valid for people to raise any sort of
objection, however vague it might be. However, a repeated assertion to
address a particular issue I believe carries with it some obligation to
make a recommendation that one can act on. Every request must certainly
not be a change proposal, but to be considered a change proposal I would
think that a specific change to a document must be specified.
As examples of *fictitious* comments that span this range
1. Your proposal is bad.
-- just a statement of preference, not actionable
2. All ontologies that use VNGB set theory are bad, I insist you use ZF.
-- maybe a valid point to discuss, but still not actionable, might not
even be theoretically possible, who knows?
3. All ontologies that use VNGB set theory are bad, I insist you use
ZF. An example of an ontology which uses ZF is http://foo.baz.org
-- better, provides a concrete example, people can discuss whether the
example is relevant and if so, use it to guide a change
4. All ontologies that use VNGB set theory are bad, I insist you use
ZF. An example of an ontology which uses ZF is http://foo.baz.org. In
your ontology at line 27, change the first axiom for Set to read (=> (...)
(...))
-- an actual change proposal. It might not be a good one, but is
concrete enough to discuss, and if it's good, it's also immediately actionable
For my part, I'll try to address every issue, because I'd like to find
solutions that allow everyone to support our ontology. Despite the
heckling my 3-d vs 4-d proposal has gotten, I hope folks will recognize the
intent of trying to solve an issue that is of great concern to a number of
people, even when not specified as a change proposal.
Adam
At 08:21 AM 8/28/2001 +0100, Tim King wrote:
>I request clarification here. As an active participant in ISO standards
>development, I would characterise that process as being issues driven. I
>can not perceive that it is possible to mandate a level of detail on the
>proposed change column of the ballot forms; the editor of the document
>has to demonstrate that the issue has been addressed through a resolution
>of the issue. Although the difference may only be subtle, is Adam right
>to claim that this IEEE activity is driven by change proposals? This
>issue begs a supplementary question: is someone collating all the ballot
>comments into an issues log that can be examined at all points in the life
>history of the development. This log should include a unique identifier
>and the eventual resolution.
>
>I see issues as being user requirements and I expect that many of us will
>be using our votes to protect our domain communities from a standard that
>fails to address the needs of the users. A subset of the participants in
>this mailing list will be capable to produce the robust solutions that are
>demanded. These people fail to address all ballot issues at peril of not
>converting NO votes into YES.
>
>Cheers,
>Tim.
>
> > -----Original Message-----
> > From: Adam Pease
> [<mailto:apease@ks.teknowledge.com>mailto:apease@ks.teknowledge.com]
> > Sent: 27 August 2001 15:36
> > To: John F. Sowa
> > Cc: West, Matthew R SITI-GREA-UK; Yang Yun;
> > standard-upper-ontology@ieee.org; phayes@ai.uwf.edu
> > Subject: Re: SUO: 2000-7-26 example
> >
> >
> >
> > John,
> >
> > At 10:06 PM 8/26/2001 -0400, John F. Sowa wrote:
> > >Adam,
> > >
> > >My first change proposal is the same one that I recommended
> > >at the SUO workshop at IJCAI.
> > >
> > >AP> Absolutely, what is your change proposal (stated as specific
> > > > changes to the document)?
> > >
> > >Merge Robert Kent's IFF document with the SUMO document.
> >
> > This is a reasonable recommendation for a task (albeit a big
> > one), but it
> > is certainly not a change proposal. Take KIF as an example.
> > "Include
> > default reasoning in KIF" would be a recommendation for a
> > task, and, like
> > your recommendation for IFF and SUMO, one the success of which is not
> > guaranteed. On the other hand
> >
> > change the BNF grammar in line 21, page 15 from
> >
> > foo ::= bar | baz
> >
> > to
> >
> > foo :: bar* | baz | bif
> >
> > would be a concrete change proposal.
> >
> >
> > >The following point is fine, but my notion of feature is much more
> > >than just an axiom or two here or there.
> >
> > It could be a substantial change or even a completely new
> > proposal, but
> > unless we get down to specifics it's very difficult to make progress.
> >
> > >AP> I think my stance is simply the as you say that "anyone
> > who proposes
> > >a new
> > > > feature has a responsibility
> > > > to explain the importance of that feature and why it should be
> > > > adopted".
> > >
> > >I am writing up a more detailed explanation of the lattice of
> > >theories and its implication for the ontology standards.
> > >
> > > > I'm not defending SUMO, just requesting that whoever says
> > > > something is wrong with it must say precisely what it wrong
> > > > (in terms of its terms or axioms), and provide a revision
> > > > to the document.
> > >
> > >That parenthetical comment is the one I most object to. As I have
> > >said from day one of the SUO effort, I believe that the methodology
> > >for organizing the theories and axioms is far more important than
> > >any particular axioms.
> >
> > Well, sure methodology is important, but we're not producing
> > a methodology
> > as the standard. Even if we accept that the methodology must be done
> > first, at some point, one still has to specify terms and
> > axioms. Anyway, I
> > fear we've been over that ground before.
> >
> > >The urge to write axioms is very much like a programmer's desire
> > >to write code when the overall architecture hasn't been done.
> > >I believe that there is an important amount of work that has been
> > >done on how to organize an ontology in the past 10 years and none
> > >of it has been reflected in SUMO.
> >
> > "None" is rather strong, especially since many of the
> > ontologies, if not
> > all, that we've incorporated, we're built in the last 10 years.
> >
> > >I've been sketching out some
> > >of the ideas in my comments, and Robert K. has been proposing them
> > >in his IFF document. I realize that IFF is at too high a level
> > >of abstraction for the underlying ideas to become apparent, but
> > >I'm trying to bridge that gap.
> >
> > I see no way to bridge that gap other than actually writing
> > or changing the
> > axioms in IFF and SUMO to show how they can be made
> > compatible. Explanations will help but, just as we can't
> > create a model
> > theory for KIF by English prose alone, so is prose also
> > insufficient in
> > this case.
> >
> > Adam
> >
> > >John Sowa
> >
> > Adam Pease
> > Teknowledge
> > (650) 424-0500 x571
> >
Adam Pease
Teknowledge
(650) 424-0500 x571