Ed,
Replies are interpolated below, but one point needs emphasis up front: (01)
[EB] > Pat mistakes creating a consortium of 100 participants whose
ostensible
> purpose is to create a "foundation ontology" for "creating communities
> of practice". It is very likely that these people will represent 30
> different communities of interest, and at least 3 of them will be pure
> philosopher gangs who are all about consistency with varios sacred
> theories and disagree with each other. Communities of practice _use_
> an
> ontology and derive benefit from it. How will these folk use the
> foundation ontology and derive benefit from it? What will it enable
> that they aren't already doing in their application community?
>
> But perhaps I misunderstand. The 100 people can't possibly be drawn
> from existing communities of practice of any significance, because such
> communities don't exist now, and won't exist for several decades.
> Right?
>
No, there is no mistaking what exists for what can be created. As I
visualize such a project, funded participants will commit to actually using
the foundation ontology they develop in some non-trivial application
(meaning, not just data in , data out), over a period of at least two years,
and provide a report on how well it performed, with suggestions for needed
improvements. Projects that demonstrate interoperation of separately
developed applications will be most desirable. There can also be non-funded
participants who need not make such a commitment, and several of the
participants can combine to develop one application. The application will
have to be sufficiently open-source (some non-essential parts may be
proprietary) that the public can test it and learn from it. One possible
application might be database federation. At least one application should
be natural language understanding which includes the ability to query an
FO-based knowledge base using natural language. This is what I mean by
creating a user community. Other requirements and the details would have to
be developed by the participants themselves, before the development phase of
the project starts. (02)
More below. (03)
Pat (04)
Patrick Cassidy
MICRA, Inc.
908-561-3416
cell: 908-565-4053
cassidy@xxxxxxxxx (05)
> -----Original Message-----
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-
> bounces@xxxxxxxxxxxxxxxx] On Behalf Of Ed Barkmeyer
> Sent: Wednesday, February 11, 2009 12:26 PM
> To: [ontolog-forum]
> Subject: Re: [ontolog-forum] standard ontology
>
> Patrick Cassidy wrote:
> > My difficulty with just waiting until some "large and economically
> important
> > community of practice" evolves is, that on the basis of the
> experience of
> > the past fifteen years, this process could take several decades, or
> longer.
>
> Do I understand Pat to say that there won't be significant commercial
> interests that depend on ontologies for "several decades"? That would
> argue strongly against the significance of "foundation ontologies" for
> the same time period.
>
[[PC]] There are existing applications, but all of those that I am aware of
use a common ontology locally within a single group or specialized
community. In the absence of a widely accepted common FO, this situation is
likely to persist for quite a while. Since the character of the common FO
that is not shared by other ontologies is its ability to support translation
among multiple alternative knowledge representations (ontologies or RDBs), a
user community that takes advantage of such a common FO *cannot* develop
until the common FO exists. It is certainly possible that one of the
existing upper ontologies could be adopted by several groups as their basis
for interoperability, but thus far there does not appear to be any
significant movement in that direction for any of the existing UOs. It's
just a lot easier for any single group or small number of groups that want
to exchange information to develop their own local FO ontology rather than
try to create an FO that can do translation - there is no advantage to do
that, if no wide user community already exists. It may happen eventually,
but the evidence suggests that this process is going to take a lot of time.
Creating a user community of significant size and diversity by funding such
a community is the way to compress the time it would otherwise take for a
community of that size to evolve. Whatever the savings there would be for
using a common FO, those saving will accrue more rapidly if the pace of
creation of the broad user community could be accelerated. (06)
> > If here is any truth to the estimate that lack of semantic
> interoperability
> > costs the country $100 billion per year in lost productivity, then it
> > appears to me to be foolishness to let a few trillion dollars of
> losses
> > accumulate rather than to attack the problem directly by creating
> that
> > "community of practice" by funding its creation.
>
> I agree with this, at least in principle. Where we can see value in
> creating ontologies for use in various interactions in some domain, it
> is appropriate to fund their development. The biosciences and medical
> communities have done a great deal in that very regard. And they have
> expanding communities of practice. We NIST see value in constructing
> "supply chain" and "logistics" ontologies for improving semantic
> interoperability in software involved in international trade; and we
> are
> trying to support that. And there are other such activities in
> agriculture, finance, military and "intelligence".
>
> But most of those activities have not found "upper ontologies" useful
> in
> constructing their domain ontologies.
[[PC]] No, I'm not surprised at that. Most specialized applications that I
know of (other than NLU) can function well with small ontologies or
ontologies that have few of the more abstract concepts found in the upper
ontologies, so it is a pointless waste of time to try to learn how to use
such complex ontologies, **unless one wants to communicate information with
other groups that are using that ontology**. The common FO is
indispensable **only for communication among separately developed
ontologies**. As a result, there is a chicken-and-egg situation where the
absence of a widely used FO inhibits development of ontology-based
applications that could benefit from reuse of information from a wide
community. Kick-starting a community of users is a tactic to accelerate
development of a real community of users that have practical applications
that communicate conceptual information among them.
This is one major difference between the project I am suggesting, and a
typical standards-development project. The latter usually occurs after a
community of users that want to interoperate already have applications in
place. The FO project is intended to help develop such a community, which
does not yet exist. The purpose of accelerating the development of such a
community is the anticipation that powerful applications will be developed
more quickly, and the financial benefits (and reduction of financial losses)
will occur more quickly.
If one doesn't believe that ontologies will in fact be able to support
more powerful or more efficient applications, of course, this argument is
meaningless. I do believe it because I have seen small demos that indicate
to me that truly powerful applications can be built that will depend on
ontologies for their effectiveness; this is particularly true of
applications that need to interoperate with widely dispersed systems
developed by multiple developers. (07)
> Our experience is that there is
> no disagreement in the understanding that an engine block is a physical
> object that is counted, while oil is a physical substance that is sold
> in volumes or weights. The disagreement is in whether a shipment
> becomes two shipments when it is split at a consolidation center, and
> in
> exactly who has which responsibilities in "change of ownership at the
> dock", and in which actual event determines the date that starts the
> clock on a 30-day forfeiture clause. The upper ontologies are directed
> toward concerns of the former kinds, which are _not_ semantic
> interoperability issues, but not with the latter kinds, which _are_
> semantic interoperability issues.
>
> > That is the essence of the
> > proposal that we fund a consortium of 100 or so participants who will
> > develop, and then test in their own applications, some foundation
> ontology
> > suitable to all of them. The sooner such a community does develop,
> the
> > sooner the benefits of a common ontology can begin to be felt.
>
> Pat mistakes creating a consortium of 100 participants whose ostensible
> purpose is to create a "foundation ontology" for "creating communities
> of practice". It is very likely that these people will represent 30
> different communities of interest, and at least 3 of them will be pure
> philosopher gangs who are all about consistency with varios sacred
> theories and disagree with each other. Communities of practice _use_
> an ontology and derive benefit from it. How will these folk use the
> foundation ontology and derive benefit from it? What will it enable
> that they aren't already doing in their application community?
>
> But perhaps I misunderstand. The 100 people can't possibly be drawn
> from existing communities of practice of any significance, because such
> communities don't exist now, and won't exist for several decades.
> Right?
>
> > One might argue with the estimate of losses due to semantic
> interoperability
>
> No one argues with the losses. The question is: How does creating a
> "foundation ontology" by funding 100 people reduce those losses? and by
> how much?
> (08)
[[PC]] The maximum benefits will only occur after the FO has been widely
adopted as the basis for semantic interoperability. The *gamble* in this
project is that the FO developed may not be widely used enough to generate
the maximum economic benefit(Pat Hayes's expectation). But since the
potential benefits are so large, adoption by even a small part of the
potential user community will result in savings that justify the cost. If
the maximum potential savings are 100B/yr and the cost is 30M, then use in
only 1/3,000th of the potential uses would generate a return of 100% per
year, which is way, way larger than the return that is required to justify a
Corp of Engineers flood-control project. It seems to me to be a gamble very
much worth taking. (09)
And even in the worst case, if the FO fails to attract a sustainable user
community, we will still have learned a great deal that we don't already
know about what kind of FO is useful for what kind of applications. (010)
> We have created 5 or 6 "upper" or "foundation" ontologies, without yet
> creating significant communities of practice for them, or seeing any
> reduction in the "semantic interoperability" cost. Like XML and
> Webservices, they don't of themselves solve any "semantic
> interoperability" problem. But unlike XML and Webservices, they don't
> solve any other interoperability problem, either.
>
> It is probably not worth spending 50M$ on SETI trying to contact an
> alien civilization that has already solved the problem and could teach
> us how. Is funding the 100 people more likely to produce a solution?
>
> > Waiting is not cost-effective. Why lose all that time?
>
> As one who has spent 7 years of his professional life developing
> international standards that no one ever really used, I can say that
> making another standard for no known user community is wasting the real
> time of real people -- truly losing that time. It is not what anyone
> with any real competence should be doing.
>
[[PC]] Yes, of course. The project will **create** the user community that
adhere to the standard. I do not anticipate that the first version will be
a formal standard. It will be tested and further developed over several
years by the user community supported by the project funds - and perhaps
also by non-funded participants, or commercial firms, or academics funded
independently to do work that uses the FO. If it does find wide use and
appear to be likely to persist in utility, at that point a formal standard
project may well be worthwhile. (011)
> But it may be beneficial to many communities of practice to spend the
> money to put 100 selected individuals in a room to develop a worthless
> standard, so that they stop interfering with worthwhile efforts. ;-)
> (012)
[[PC]] What worthwhile efforts do you have in mind that you think are likely
to solve the semantic interoperability problem? It seems to me that much
money and effort has already been expended, with no evidence that the goal
is in sight. (013)
> -Ed
>
> --
> Edward J. Barkmeyer Email: edbark@xxxxxxxx
> National Institute of Standards & Technology
> Manufacturing Systems Integration Division
> 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528
> Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694
>
> "The opinions expressed above do not reflect consensus of NIST,
> and have not been reviewed by any Government authority."
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
> To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
> (014)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (015)
|