[Top] [All Lists]

Re: [ontolog-forum] Integrating Semantic Systems

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Thu, 10 Jun 2010 09:16:25 -0400
Message-id: <4C10E5A9.30700@xxxxxxxxxxx>
Azamat,    (01)

I mostly agree with your first 6 points (in the copy below).
If we got down to details, I'm sure there that we would find
many debatable issues.  But at the level of one sentence per
point, it's vague enough that we can agree.    (02)

But the seventh point raises some serious questions:    (03)

AA> 7. General Semantic Interoperability Standard implies Standard
> Foundation Ontology.    (04)

The three basic questions are about the definitions of the subject,
verb, and object of that sentence:    (05)

  1. What does the word 'general' mean?  Does it mean for all
     possible systems in the world?  Some systems?  Just those
     systems that are required to interoperate?    (06)

  2. What does the verb 'implies' mean?  Does it mean that the
     object is a requirement or prerequisite for the subject?    (07)

  3. What does the object mean?  A single standard for everything?
     A family of standards?  An underspecified standard that has to
     be made more detailed for each area of application?    (08)

I'll discuss these points in terms of the slides. I posted a new
version today, but the numbering and most of the content is the
same as in earlier versions.  For the latest version, see    (09)

    Integrating Semantic Systems    (010)

Slide 4:    (011)

  * Legacy software:  Half a trillion lines of code.
  * Percentage that is tagged with semantics:  Slightly over 0%    (012)

That code is not going away for a long time, and new code that
uses the same conventions (and implicit semantics) is being
written every day.  Any new systems must interoperate with it
for a very long time.    (013)

Could a standard ontology support interoperability with legacy
systems?  Would it help, hurt, or do nothing?  How and why?    (014)

Slide 7:    (015)

Why semantic technology isn't used more widely:
  * Too much time, effort, and training to specify semantics.
  * Delayed implementation without any obvious benefit.
  * Difficulty of sharing semantics among different tools,
    especially tools designed for different methodologies.    (016)

Note that most developers (and their managers) don't see much,
if any, advantage from using currently available semantic tools.    (017)

The Semantic Web has been pushing the idea of semantics for over
a dozen years, and researchers in AI and databases designed and
implemented even more advanced technology over 30 years ago.    (018)

Yet the amount of explicit semantics keeps going down.  OWL has
much less semantics than Cyc, RDFS has less semantics than OWL,
and RDFa has less semantics than RDFS.  In fact, Linked Open Data
(LOD) supports spreadsheets and files with Comma Separated Values,
which have no explicit semantics of any kind.    (019)

Note that OpenCyc, a very large ontology, has been freely available
for several years.  A few researchers have downloaded it, but there
is no stampede toward adopting it or any other large ontology.    (020)

Slides 13 to 15:    (021)

$70 million was invested in Cyc over the first 20 years (from
1984 to 2004), but there was very little ROI.    (022)

Since then, Cyc has focused on application development, and
the results have been more promising.  But Cyc was rejected
as a basis or even guide for the Semantic Web.  (R. V. Guha,
for example, had been an associate director of Cyc, but he
ignored it when he teamed up with Tim Bray to define RDF.)    (023)

Slides 16 to 23:    (024)

The work on conceptual schemas in the database community
has had some success in developing useful tools, but they
were ignored by the developers of the Semantic Web.    (025)

Why would anybody pay attention to yet another proposal for
an ontology?  Or for a proposal to standardize something
whose commercial value is still questionable?    (026)

Slides 61 to 81 about the SIO Project (Sharing and Integrating
Ontologies) discuss the above issues and suggest solutions.    (027)

Slides 64 to 67 focus on the interfaces.  They make the point
that the interface between systems determines what must be
consistent.  Their internal ontologies and representations
may be very different, but they can share data about areas
of agreement.    (028)

Slides 68 to 79 discuss issues about logic, reasoning, and
tools to support them.    (029)

Slides 80 and 81 discuss directions for the SIO project.    (030)

Finally, slides 82 to 99 discuss future possibilities.    (031)

As I said before, I'm not against the idea of a standard
ontology, but I view it as a goal rather than a starting
point.    (032)

One reason why I like the SIO approach is that it focuses
on the tools and methodologies for combining and relating
ontologies and parts of ontologies.  Any proposals for a
good starting ontology can be added to the OOR, and the
best parts can be adopted and integrated into the hierarchy.
Meanwhile, the tools can be used to relate and combine the
parts of any or every ontology in a systematic way.    (033)

John    (034)

-------- Original Message --------
Subject: Re: [ontolog-forum] Integrating Semantic Systems
Date: Tue, 8 Jun 2010 22:02:17 +0300
From: AzamatAbdoullaev <abdoul@xxxxxxxxxxxxxx>
To: [ontolog-forum]  <ontolog-forum@xxxxxxxxxxxxxxxx>
CC: John F. Sowa <sowa@xxxxxxxxxxx>    (035)

John,    (036)

Read the opening pages, but only glancing at the whole document.
Some general comments and suggestions about key notions, semantics,
semantic systems, interoperability, information traffic, which
might be useful to better your good content:    (037)

1. Interoperability ranges from railways to medical technologies to
telecommunication systems to business processes to software systems and to
human organizations.
2. Semantics is concerned with the semantic interoperability and integration
of all communicating entities.
3. Machine semantics is about computable semantic interoperability among
computer systems.
4. Semantic systems are about (intelligent) computing machines using general
semantic interoperability model.
5. Semantic interoperability implies syntactic interoperability dealing with
data formats and common protocols, and KR languages (as XML, SQL, RDF, OWL,
and other FOL-like standards)
6. Information traffic is going to YB level
"In 2008, Americans ...Consumption totaled 3.6 zettabytes and 10,845
trillion words, corresponding to 100,500 words and 34 gigabytes for an
average person on an average day." How Much Information? 2009 Report on
American Consumers.
7. General Semantic Interoperability Standard implies Standard Foundation
Ontology.    (038)

Last not least, the Related Reading is good to widen to all key directions    (039)

Azamat Abdoullaev    (040)

http://www.standardontology.com    (041)

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    (042)

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