Re: [ontology-summit] Reusability and Interoperability (was Proceedings:

[Gary Berg-Cross <gbergcross@xxxxxxxxx>]

The Basic point that John makes about reuse is one that I agree with and is important to this discussion of

reusing semantic content in Big Data, Semantic Web and LOD applications:  

Different kinds of applications have different requirements for ontology.  There is no such thing as a common

definition of reusability or interoperability that can cover

all the versions.

One useful task for this track discussion is to provide illuminating examples of these different requirements.

There is

 perhaps some  

degree of generality that might be found within particular types of applications.

John>

For ontologies, an underspecified ontology such as Schema.org can

be reused in many applications on different platforms.  Some OWL
ontologies are easy to move -- provided that they don't use anything
beyond Aristotle -- eg, the GoodRelations ontology, which was moved
to Schema.org.

Here I wonder if we would be selling our experience and understanding short.

Haven't we made progress on understanding several areas of semantic relations that

can be reused.  Distinctions among different types of Part relations come to mind and

are post-Aristotle.

In addition I would suggest that we can leverage some existing conceptual structures (apologies for

reusing your phrase lightly of discussion purposes) such as found in domains to connect concepts.

The link of diagnosis-illness-genes mentioned by Michel is perhaps one obvious idea to leverage.

The junk yard metaphor of piece reuse  is not one that had occurred to me and

I'm not sure how far we would take this in terms of granular parts and what I have expressed above

suggests rummaging around for useful grains in perhaps a semi-organized antique store.  

I'm sure that others will have comments on these topics as well as other parts of John's comments and

this would be useful for the Track and Summit as a whole.  So have at it.

Gary Berg-Cross, Ph.D.  

gbergcross@xxxxxxxxx     

http://ontolog.cim3.net/cgi-bin/wiki.pl?GaryBergCross

NSF INTEROP Project  

http://www.nsf.gov/awardsearch/showAward.do?AwardNumber=0955816

SOCoP Executive Secretary

Knowledge Strategies    

Potomac, MD

240-426-0770

On Mon, Mar 10, 2014 at 9:20 AM, John F Sowa <sowa@xxxxxxxxxxx> wrote:

Peter,

I agree:

> We had another great session on Thursday (Mar-6) ... Track-A:
> Common Reusable Semantic Content-II - Experiences in Knowledge
> Sharing: Lessons from research and experience in Big Data,
> Linked Data and Semantic Web Applications.

But I changed the subject line of this thread to reflect some
issues that came up in the questions (both oral and typed).

Basic point:  Different kinds of applications have different
requirements for ontology.  There is no such thing as a common
definition of reusability or interoperability that can cover
all the versions.

In fact, there is no common definition of 'reuse' or 'interoperable'
that covers all the ways that artifacts of any kind -- physical
or computational -- can be used or reused.

Physically, there is very little reuse, except as spare parts from
junkyards.  And they're mainly used to replace identical or nearly
identical parts in objects for which the manufacturers *planned*
in advance to have a modest amount of reusability.

Computationally, the situation is similar:  parts can only be reused
on the same platform.  You can build parts on a general platform,
such as Posix, and convert them to other platforms.  But if you build
on a very specialized platform (eg, Windows or Apple), you can only
reuse the parts on variations of the same platform.

You can move data across platforms.  But the more structure there
is in the data, the harder it is to move.  Just imagine the problems
of moving an SQL database from Oracle to DB2.  Even moving a PDF file
to a DOC or HTML file is nontrivial.  You either get a very buggy
version, or you get a clean version by brute force:  convert each
PDF page to JPG and embed it in a DOC or HTML wrapper.

For ontologies, an underspecified ontology such as Schema.org can
be reused in many applications on different platforms.  Some OWL
ontologies are easy to move -- provided that they don't use anything
beyond Aristotle -- eg, the GoodRelations ontology, which was moved
to Schema.org.

Any ontology that has more detail can only be reused within the
platform it was designed for.  RDF and RDFS are so limited that
anything defined in them can be very underspecified -- and
therefore "easier" to reuse.

But people can and do use RDF and RDFS with a large amount of local
"conventions".  But those conventions are an informal ontology whose
details are specified only in the comments.  The problems of reusing
such data are the same as the problems of reusing any structured data
with some specialized, idiosyncratic, poorly defined structure.

John
_____________________________________________________________________

For the slides, audio, and chat from the session, see the archives:
http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2014_03_06

_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014
Community Portal: http://ontolog.cim3.net/wiki/

_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/   
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/  
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014  
Community Portal: http://ontolog.cim3.net/wiki/     (01)