Todd, Leo, and Chris, (01)
I also sympathize with your comments. The *best* software development
tools and methodologies of the past 30+ years had a strong ontology
component. But they did not use the word 'ontology'. (02)
The best examples are the conceptual schema efforts from the 1978
ANSI/SPARC technical report, the 1987 ISO TR, and the 1999 ISO TR.
But note that all three of those projects started out as *standards*
projects, but they all terminated as TRs. I don't believe that is
a coincidence. That is also a warning sign for the Ontology Summit. (03)
As an example of a good introduction to ontology development, which
I have recommended and Pat supported, is a textbook on Aristotle's
ontology and syllogisms. Sister Miriam Joseph taught it to freshman
English majors at St. Mary's college from the 1930s to 1960: (04)
Joseph, Sister Miriam (1937) The Trivium: The Liberal Arts of Logic,
Grammar, and Rhetoric, third edition 1948, reprinted by Paul Dry
Books, 2002. (05)
Note that it was reprinted in 2002 (after being revived by some of
Sister Miriam's former students), and it is still one of the top
sellers for books on language and logic at Amazon. (That's partly
because it's very well written, and partly because very few people
read books on language and logic.) (06)
TJS
> More importantly than lifecycle phases is the need to show
> the software and systems engineering communities and their
> management that employing ontologies and semantic technologies
> is not procedurally different that what is currently done
> (i.e., it does not incur significant risks). (07)
I very strongly agree. Unfortunately, the current tools for
ontology are not integrated with the software development tools.
For that purpose, UML diagrams are far superior. (08)
Leo
> an ontology lifecycle regimen is absolutely needed. And it largely
> uses but expands on the software development cycle, such as it is,
> i.e., is rooted also in software/systems engineering, (09)
Suggested regimen: Sister Miriam's students, for example, could start
using UML for ER-diagrams, type hierarchies, and activity diagrams
tomorrow. For the next step, they could integrate them with rule-based
tools and database tools that had a good front end. But requiring them
to learn RDF, OWL, and SPARQL would be a huge step *backwards* . (010)
Leo
> Web ontologies are often lightweight and are often a hodge-podge (011)
Sister Miriam taught her students the difference between types and
instances. She also taught them conceptual analysis, how to relate
English grammar to logic, and how to detect fallacious reasoning. (012)
Leo
> I think ontology lifecycle should explicitly figure into ontology evaluation. (013)
That point was strongly emphasized in the conceptual schema TRs.
But the academics who jumped on the DAML bandwagon had no understanding
or even awareness of those issues. They didn't reinvent the wheel
-- they flattened it. (014)
CP
> there seems a trend in IS development to increasingly ignore the areas
> in which ontology might contribute to development and focus on the
> physical implementation issues. (015)
Yes, indeed! In the 1960s, Knuth and others said "Make it run before
you make it faster." In the 1970s and '80s, the conceptual schema
developers fought a long, hard battle to get software developers to
think in terms of ontology instead of bits. (016)
One of the early promoters of the conceptual schema was Tom Steele
from AT&T. His slogan was "The only logical notation for the
conceptual schema is logic." Those guys had a better understanding
of logic and ontology than most members of the W3C today! (017)
CP
> I'm not sure looking at Cyc (as Pat suggests) could really help,
> as this is not a typical information system (018)
I certainly agree with the second line. In the early 1990s, I was
trying to convince Lenat that they needed to do more work on application
development. But they didn't shift their focus until DARPA cut their
budget in 2005. They are now doing the kind of work on applications
that they should have been doing in the 1990s. (019)
CP
> I would agree with you (Leo) that "an ontology lifecycle regimen is
> absolutely needed", the problem is what this should look like. (020)
Yes, indeed. If you don't have an answer to the second clause, it's
a sign that you need to do more work *before* you recommend anything. (021)
CP
> the whole issue of methodologies for legacy modernisation is a vexed one
> - there are next to none despite this being a growing requirement (022)
It's not a growing requirement -- it's the *primary* requirement.
Anything that cannot interoperate with the software that runs the
world economy is a *toy* . (023)
CP
> I suspect that questions about whether the whole process can be
> rationalised have not really been looked at. (024)
Yes! Yes!! Yes!!! (025)
John (026)
_________________________________________________________________
Committee Msg Archives: http://ontolog.cim3.net/forum/ontology-summit-advisors/
Subscriber Config:
http://ontolog.cim3.net/mailman/listinfo/ontology-summit-advisors/
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2013/
Community Discussion: : http://ontolog.cim3.net/forum/ontology-summit/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2013
Community Portal: http://ontolog.cim3.net/wiki/ (027)
|