SUO: Resolution of Ballot Comments
SUO,
Pierluigi will be away for a couple weeks, so I will lead the
resolution of ballot comments, with the goal of crafting a ballot
question
that maximizes consensus. Plenty of time will be provided for members to
comment prior to release of any further ballots.
Jumping to the bottom line, I offer (based on analysis of ballot
comments) the following draft revised ballot question:
Ballot Question: Should the IEEE SUO WG commence work on OpenCyc?
Review of ballot comments: (See actual comments below)
1. Many comments (too many to cite) suggested Opencyc, CycL, and CycML be
split into separate actions. As such, the above ballot question
addresses
only OpenCyc. CycL and CycML can be proposed later, if appropriate.
2. Ian Niles raised the issue of 'control of the document,' which gets
down
to copyright. I will summarize where we stand at this point, but give me
a few days.
3. Pat Cassidy asked, "I would like more information on how
OpenCyC will be related to SUMO or any other candidate -- will an attempt
be made to create a merged SUO, or will the OpenCyc candidate merely
mirror the developments at Cycorp??"
In a related comment, Robert Kent asked for wording in the ballot
on
'collaboration with the other starter documents.'
This is an issue for Pierluigi Miraglia or John De Oliveira to
respond to. It may, or may not, require wording in the ballot. But
remember, the ballot belongs to the group and should represent group
consensus, not the intent of one party.
4. Adam Pease asked for the ontology to be offered in a
single
text file. Adam, do you mean you'd like to see this before this vote
takes
place, or as the final standard?
Please discuss any aspect of these issues, or any other issues
relating to this ballot.
Jim Schoening
=====Ballot comments: (Please post any additional comments to the
Standard-upper-ontology list.) =================
I vote "no".
This to be interpreted as supporting the split of the motion into
separate bits.
Matthew West
Email: matthew.r.west@is.shell.com
=====================
I vote NO on the ballot question below for two reasons. First, as
many on the list have pointed out, OpenCyc needs to be cleanly separated
from CycL and CycML. Second, it's not clear to me that the SUO WG would
have sufficient control over the development of extensions to and
modifications of OpenCyc content. As I understand it, Cycorp intends to
eventually move over a large amount of content in full Cyc to OpenCyc.
However, it is possible that some of this content will be rejected by the
group or that the group will recommend changes to content that is already
part of OpenCyc. In these two cases, it is not clear to me that Cycorp
would be willing to acquiesce to the will of the group. If Cycorp is
willing to surrender control of the document to the group, there should,
I
think, be a formal statement to that effect from the company.
-Ian Niles
iniles@teknowledge.com
=======================
I vote NO on the below ballot question.
I am unhappy with the wording of the present motion, since (whether
intended
or not) the adoption of the OpenCyc Ontology seems to be coupled with
adoption of CycL and CycML. I vote NO in order to dispose of the current
motion. However, I am very much open to a new, rephrased motion to the
effect that the IEEE SUO WG should commence work on the OpenCyc Ontology.
However, I see no need now or in the future for CycML.
From my perspective, it is OK for any candidate ontology, OpenCyc, SUMO,
etc., to be expressed in any language having appropriate features of
many-sorted first-order logic. There are two ways to handle candidate
ontologies that are expressed in different languages.
* The first way is to collaborate with the Common Logic group.
* The second way is to define IFF Input/Output portals for the
candidate
ontology.
IFF I/O portals make the candidate ontology compliant with and provide a
"translation manual" into the "standard" framework (IFF). For more on
this
see the two example input portals (traditional logic, conceptual graph)
linked from the IFF Model Theory Ontology
<http://suo.ieee.org/IFF/versions/20020515/IFFModelTheoryOntology/Introdu
cti
on.htm>.
I also recommend that the new proposal for OpenCyc have some wording that
indicates a willingness to collaborate with the developers of the other
two
candidates, SUMO and IFF, in order to work toward a convergence that
incorporates the best features of all three. In particular, work needs to
be
done on the alignment and unification of interesting submodules of these
ontologies under the auspices of the "concept lattice of all theories" as
provided by the IFF.
Robert E. Kent
rekent@ontologos.org
===========
My vote is NO on this motion. I do not feel that this is focussed
enough at this time to result in any definite work. I do welcome
the open Cyc to SUO in some form, however.
Best,
John Velman
<velman@cox.net>
=======================
I vote YES.
I know I don't have to comment on a "yes" vote, but, for the record, my
reason for this vote is to get another source of content, and another
team
involved in the process. Recent investigations of SUMO and interaction
with the SUMO team have led to the unhappy but inescapable conclusion
that
there is no support for, or agreement with, what I consider to be the
primary purpose of an upper ontology. I'm hoping that an alternative
source will provide an avenue in the direction that I have expected this
work would eventually take.
Doug McDavid
mcdavid@us.ibm.com
=======================
I vote NO, on the basis that the group feels that there has not been
enough
discussion of the motion, and that a revised motion can easily be
resubmitted once there has been more discussion and more of a consensus
reached. It seems to me that this consensus is more important that losing
a
few months may be for the development of the SUO.
Chris
chris_partridge@csi.com
=======================
I vote NO on the motion to accept OpenCyc as a candidate
document.
I think many concepts from OpenCyc would be very valuable
as a component of an SUO, but I would like to see the
proposal made for the OpenCyc conceptual content alone, without
the language. Also, I would like more information on how
OpenCyC will be related to SUMO or any other candidate --
will an attempt be made to create a merged SUO, or will the
OpenCyc candidate merely mirror the developments at Cycorp??
Pat Cassidy
pcassidy@bellatlantic.net
===================
I vote NO on the OpenCyc proposal.
I will vote Yes when it is unbundled from CycL and such.
John A. Thompson
The Boeing Company
<john.a.thompson@boeing.com>
=====================
I reluctantly vote NO.
Since Pierluigi has indicated his willingness to submit a separate
motion of just the OpenCyc ontology, I'd like to give him the opportunity
to do that. I would also want the ontology to be offered as a single
text
file in the CycL language (if CycL is the language OpenCyc folks intend
to
use rather than KIF or CL) so as to separate the ontology itself from any
tools that may facilitate its inspection.
Adam Pease
<apease@ks.teknowledge.com>
======================
I also vote NO on the motion to accept OpenCyc as a candidate
document for a standard.
As I said in my earlier notes, I would very much like to see
a standards project that would contribute the content of the
OpenCyc ontology to the SUO efforts, but the motion should be
reworded, as we have already discussed.
John Sowa
<sowa@bestweb.net>
=============
I vote ABSTAIN to the motion of OpenCyc.
Although I think its good to work on OpenCyc to use parts
of it (especially its capability for logical inference and reasoning)
for the ontology standard. And although I, opposite to
some other SUO members believe that the portion which
should be standardized is the language and inference
mechanisms for the ontology not its contents, But regards
to the recent discussion in the list about the rewording the
motion and according to Pierluigi's mail which showed he
also agrees to change the ballot wording and break it to
three separate ballots, I vote Abstain.
Regards
Mehrnoush Shamsfard shams@pnu.ac.ir
=====================================
I vote 'no' on the Open Cyc motion, mostly for reasons I alluded to in a
note to the SUO list (Re: SUO: OpenCyc motion) on June 1.
I think that the ontology content of OpenCyc is worthy of SUO
consideration
and I hope that a "rider-free" motion along such lines is submitted
should
the present OpenCyc motion fail.
best regards,
Mike Pool mpool@iet.com
=================================
I feel that the ballot was not suitable and the procedure is flawed. I
can find as many good reason to vote yes on the basis that OpenCyc has a
relevant fragment to the puprose of this work or to vote no on the basis
that the scope of the current submission is too broad and mostly
irrelevant. I am worried about the standard of quality that this group
sets for itself.
I abstain.
Sincerely,
Pierre Grenon pierre.grenon@ifomis.uni-leipzig.de
===========================================
I vote NO on the motion to accept OpenCyc, CycML, and CycL as
candidate standards. Without disregarding the historical importance
of the Cyc endeavour, and the practical value of some of its results,
I don't think it can be envisaged as a "standard", even a candidate
one - see Section 3.44 of our 2001 book on "Intelligent Database
Systems".
More in general, I am not convinced that the present,
informal discussion about some (often minor) aspects of pre-existing
industrial proposals (SUMO, Cyc) could ever produce a valuable
standard. To my knowledge, proposing a standard implies the creation
of ad-hoc committees and working groups with very precise goals and
schedulings, able to organise face-to-face meetings, producing
reports etc. This implies in turn the disponibility of some money, as
it was already claimed (without success) at the very beginning of
this group.
G.P. Zarri zarri@ivry.cnrs.fr
===============
With respect to the current ballot on OpenCyc:
I vote No, on the grounds that the current tripartite motion is too broad
(as it includes, besides OpenCyc, both CycML and CycL), and too
imprecise.
I would like to see the current motion split, revised in each part, and
presented as separate motions.
Jay Halcomb jhalcomb8@attbi.com
=======================================
________________________________________________________________
GET INTERNET ACCESS FROM JUNO!
Juno offers FREE or PREMIUM Internet access for less!
Join Juno today! For your FREE software, visit:
http://dl.www.juno.com/get/web/.