Re: SUO: Vote 2001-02: IFF Foundation Ontology
Folks,
I vote ABSTAIN, with comments.
1. Robert has created a precise and detailed document. He has done the
hard work to make a concrete proposal instead of just making criticisms or
recommendations. His proposal is stated in a formal logical language. It
addresses the requirements of the PAR Scope. For all that he is to be
commended.
2. At this stage my interpretation of the vote is just that we're saying
that the document is relevant enough to the PAR to warrant that the group
address and evaluate it. Discussion absent a specific proposal leads only
to more discussion so having a document to focus on is essential for a
standards effort and warrants more than just having one or two people work
on a document with no official acknowledgement from the group.
3. On the other hand, I do not think that an adequate case has been made
for how IFF addresses the PAR Purpose. Specifically addressing the Purpose:
a. I'm not aware of any reasoners that can actually perform tractable
inference on these constructs. Specware has tried to do this, and it has
been over 2 years since I was acquainted in any detail with the system, but
at that time it still required a great deal of human-computer collaboration
to perform proofs. I'd be happy to be made aware otherwise. In contrast,
many systems such as SNARK, OTTER, PTTP do tractable inference in first
order logic (although of course, some problems that are expressible are not
tractable).
b. For interoperability the case is stronger since Kestrel has used its
category theory tools for software interoperability. I think this is still
much harder and less mature than using first order logic for program
interoperability, but it has been done in practice at least by a very small
number of experts. Since first order logic also has a limited set of
experts in this task, the difference is only an arguable one of degree and
not kind.
c. For application areas I think the case for IFF is once again weak
although there's no inherent reason why a stronger case couldn't be made by
expanding the ontology considerably in the number of "real-world" concepts
linked into IFF existing "upper-upper" level. I'd suggest that it needs
activities similar to the QoS ontology, or the WordNet mapping that we're
doing with SUMO to make a stronger case for its relevance to applications.
So, in summary, it's a real proposal and so I can't stand in the way since
the author feels it's a suitable basis for the standard. On the other
hand, I can't give it a hearty vote of support (yet) because of the points
above.
Adam
At 10:46 PM 8/9/2001 -0400, Schoening, James R CECOM DCSC4I wrote:
>ATTN Voting Members of the SUO WG,
>
>1. Please immediately acknowledge receipt of this email letter ballot. If I
>do not receive such acknowledgement within 5 days, I will contact you again.
>We want to make sure all voting members receive this ballot.
>
>2. This message is to ballot the question, as proposed by Robert Kent:
>
>"Should the IEEE P1600.1 Standard Upper Ontology Working Group commence work
>on the IFF Foundation Ontology version 1.0 [July 20, 2001] posted at
>http://suo.ieee.org/Kent-IFF.pdf, with the intent of developing it into the
>final SUO document?
>Note 1: See the background information below for more details.
>Note 2: This may be one of several candidate documents, any one of which may
>become the final SUO document, or may be combined and aligned into the final
>SUO document via the consensus building process. No one candidate document
>will have preferential status relative to the others."
>
>3. By Thursday, August 30, 2001, please vote YES (comments optional),
>NO(comments required), or ABSTAIN (comments optional). Votes should be sent
>directly to me (to reduce list traffic), but please post your comments
>directly to the list so they have a chance to be resolved. Comments sent to
>me will be forwarded to the SUO list.
>
>4. Per the SUO Chair's interpretation of Robert's Rules of Order (with input
>from members of the SUO P&P Subgroup), the following items will be followed:
>This vote will pass if it receives a majority of YES vs. NO votes. ABSTAINS
>and non-votes are not counted in determining majority. The Chair may vote to
>make or break a tie. As all voting members will have received and
>acknowledged receipt of this ballot and will therefore have an opportunity
>to vote, there will be no quorum of votes required for this motion to be
>determined. None of these items are considered to be changes to existing
>policies and procedures, only interpretations.
>
>5. Non-voting participants are welcome to express opinions and submit
>comments to the SUO list.
>
>6. Background:
>
>a. The purpose of this vote is to determine the level of consensus we have
>for focusing work on this document. In other words, we need to decide if we
>are going to go down this path or not. A working group may decide to work on
>more than one document.
>
>b. The primary purpose of this vote is NOT to propose improvements to the
>document, unless they are major enough for you to vote against focusing work
>on this document. However, such general improvement comments are welcome and
>will be resolved (but not unless and until the vote passes, since there
>would be no sense resolving comments if the group does not want to proceed
>with this document.
>
>c. If this vote passes, this document will change from being the work (and
>under the control) of a group of individuals to being the work (and under
>the control) of the SUO WG. A Technical Editor will be appointed or
>elected, who will incorporate WG-approved changes into the document. This
>does not require frequent formal votes for every change. A better approach
>for daily or weekly updates is to develop consensus by resolving objections,
>as we did when revising the Scope and Purpose. In other words, (1) the
>Technical Editor leads the discussion by processing the issues and
>suggestions; (2) the issues/suggestions identify proposed additions,
>changes, or deletions to the wording in the document; (3) the Technical
>Editor asks the WG for any objections. Items that do not have clear
>consensus would be accumulated for later, more formal processing, i.e., the
>proposer of the issue/suggestion is responsible for making a formal proposal
>(with a WG vote) for the proposed change.
>
>Jim Schoening
>Chair, SUO WG
Adam Pease
Teknowledge
(650) 424-0500 x571