[Top] [All Lists]

Re: [ontolog-forum] Past, Present, and Future of Ontology

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Fri, 29 May 2009 16:24:49 -0400
Message-id: <4A204491.1000500@xxxxxxxxxxx>
Frank,    (01)

Thank you for the reference.  I followed it to the following source:    (02)

https://www8.imperial.ac.uk/content/dav/ad/workspaces/business-school/programmes/doctoral/Technology_Transfer_Studentship_QinetiQ.pdf    (03)

The author, Erkko Autio, also has other valuable insights in other
articles.  But I'd just like to comment on the following quotation
from the passage you cited:    (04)

 > Because of the historical dominance of market-based theoretical
 > frameworks, the dynamics of innovation in value networks are
 > poorly understood.  To understand technology transfer and diffusion
 > in such environments, it is necessary to discard market-based
 > theories and replace them with institutional and structural theories.    (05)

Autio certainly does *not* downplay the importance of the market, but
he emphasizes that success in any new venture also depends on the
"‘users’, ‘suppliers’, and ‘innovators’" and their roles in the
institutions.    (06)

For example, Alta Vista and Yahoo were dominant search engines long
before Google came along.  The market was already mature when Google
was just a student project at Stanford.  What led to the success of
Google and its continued dominance was the result of many factors,
including the innovations and the talents of the entrepreneurs.
Without a market, nobody gets any rewards.  But being the first to
market doesn't ensure long-term success.    (07)

If we apply these insights to ontology, there are some hypotheses
that seem likely:    (08)

  1. There are some applications for ontology, but there is no evidence
     of a market for ontologies, as such.  Cyc was in existence for 25
     years, and they were trying to sell ontologies for almost all that
     time with very little success.  Cycorp is now making some money
     by building applications that use ontologies rather than the
     ontologies themselves.    (09)

  2. WordNet is widely used because it is free.  But most of the
     applications are research projects, not applications that make
     a significant amount of money directly from WordNet itself.    (010)

  3. OntologyWorks is a small company that builds applications
     that use ontologies, primarily by aligning relational databases.
     They started with their own upper level ontology based on Dolce,
     but they found that the upper level wasn't very useful for
     aligning independent databases.  They found that the most
     useful ontologies were the more specialized lower levels.    (011)

  4. The Japanese EDR project (Electronic Dictionary Research) began
     around the same time as Cyc and WordNet.  They developed an
     ontology of about 400,000 concepts with mappings to both English
     and Japanese.  Multibillions of yen were spent on that project,
     but there were very few users.  They charged $20,000 per copy
     for commercial users, but universities could get a much cheaper
     copy.  Stanford has a copy of EDR, but the last I heard, nobody
     at Stanford ever used it for anything, not even for a research
     project.    (012)

  5. At VivoMind, we have developed our own ontology based on freely
     available resources, and we have developed tools for extracting
     additional information from text and tailoring it for various
     applications.  If a large free ontology were available, we would
     probably find some use for it, but we're doing quite well without
     it.  As Cyc and OntologyWorks have discovered, the reasoning is
     done at the lower levels.  Lexical resources such as WordNet,
     VerbNet, and Roget's Thesaurus together with special lexicons
     for application domains are suitable for broad coverage.    (013)

For reasons like these, I believe that developing a large expensive
ontology with the coverage of Cyc or EDR would be a waste of time,
money, and highly expensive talent.    (014)

A much cheaper and more useful approach would be to develop a
lattice or hierarchy on top of some suitable repository.  Then
populate that hierarchy with all the ontologies (big and small)
that anyone might contribute.  The hierarchy would provide the
framework for relating all the ontologies that anyone might ever
develop and contribute.    (015)

This approach would not preclude the eventual development of a
large upper level, but it could get started today with a minimum
of funding -- not much more than a web site and a good programmer.
That programmer could even be a part-time volunteer.    (016)

John Sowa    (017)

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    (018)

<Prev in Thread] Current Thread [Next in Thread>