David, (01)
(The Summit is past so now I again talk big new futures, hopefully
without unduly distracting the summiteers from their respective
post-Summit follow-up projects! Not so distant futures either,
especially if there is anyone out there who would like to join the
project...) (02)
Your post to the Summit list, now at
http://ontolog.cim3.net/forum/ontology-summit/2009-04/msg00012.html, was
too much of a temptation for me, so I give some of my own answers to the
concerns you expressed in it: (03)
> Can you please add:
>
> The STEP standards community has at least the following concerns
> related to ontologies as standards:
>
> - ontology "versioning", publishing
>
> - harmonizing, reusing, subsetting ontologies (04)
I build on my post of April 2, now at
http://ontolog.cim3.net/forum/ontology-summit/2009-04/msg00000.html,
introducing my intended Summit presentation now in annotated form at
http://TheMainstream.info/RTM.html: (05)
I can't help pointing out how both those lines of yours evoke features
which are absolutely central to the whole ontologized component market
infrastructure I am working towards with the program Metaset 1.0 as the
first architecture-canonical AOS or Application Operating System. (The
plan is that peer-to-peer user AOSs or 'market agents', eventually
superseding the old-web browser and much else too, will drive the
planned new medium I have most presumptuously starting referring to as
"the Agateweb".) (06)
Concerning versions, I would urge you to read an old post of mine now at
http://ontolog.cim3.net/forum/ontolog-forum/2008-03/msg00156.html. It
expresses a view of versioning which in its features and internals is
well in advance, I believe, of any other view of versioning I have seen
elsewhere. In a nutshell, it sees versioning as a simplifying factor,
not the complicating hassle that it too often is. More than that,
however, the application lifecycle management (ALM) functionality of the
coming environment is merely yet another case of the manageability
afforded by the reflective yet practical nth-order logic that
appropriate architectural generality tends to lead to. (07)
I would in addition expand your first line with version coexistence and
migration, and ontology and application evolution. In your Summit
presentation you did mention the OMG's QVT, so you are well aware of the
need for a systematic approach. But the OMG's entire MDA is a dreadful
mish-mash compared to the more powerful yet far simpler and
diversifiable approach made possible by The Mainstream Architecture for
Common Knowledge, on which you will gradually see more appear,
opportunistically on ontolog and more systematically on
TheMainstream.info. (08)
The functionality of your second line of wishlist above is implicit in
the way my proposed component architecture provides for component- and
product-discovery and integration which are implicit and automatic
except where the application context implies user choice. At its heart
is the fine abstraction-level approach introduced in my "2nd instalment"
from this point in it:
http://ontolog.cim3.net/forum/ontolog-forum/2008-02/msg00291.html#nid016,
and further expanded on in the brief follow-up exchange with Pat Hayes. (09)
I can now add that that fine abstraction-layering is well portrayed by
the agate metaphor of the above-linked presentation. That show's slide
26, especially its 3rd bullet, adds some important underlying features
of that distributed and scalable market. Of course, well-conceived
orthogonalities greatly facilitate the process, though, once again,
that's not too hard, seeing how our conceptual knowledge as abstraction
has evolved such that orthogonalities are discovered in the process.
It's Heraclitus' logos again (I believe), as in the slideshow. (010)
Regarding your "subsetting ontologies", that is where the fine-layering
approach (_very_ difficult or cumbersome to specify and visualize in an
ontology "language", I may add) greatly facilitates the run-time
discovery and expansion of ever higher levels of common subset-ontology.
(This isn't called "The Mainstream Architecture for Common Knowledge"
for nothing, after all.) (011)
I copy from the above-cited 2nd instalment: "I would however like to
record at this point that the MRCL, or Most Refined Common Level,
between two situations will be the basis of many miracles which ordinary
people-users will perform on a daily basis." That was at
http://ontolog.cim3.net/forum/ontolog-forum/2008-02/msg00291.html#nid024.
The first bullet on slide 25 helps give some very down-to-earth content
to that confident expectation. (012)
John S. has another name for the MRCL, so do the Category Theorists, but
I forget them both... But evidently there is much active research in
this area, most of which will surely shed much useful light on this
aspect of our future, and lead to marvellous elaborations of the
functionality here, both automated and people-driven. (013)
John has also pointed out in an earlier post (which I can't find right
now) that interoperation is usually based on a much more modest level of
commonality. (That is indeed true, even though you wouldn't think so
when you look at accepted standards for EDI, which seem very prone to
being ratchetted up into some kind of superset ... until fatigue sets in
and the standard is pegged where it is.) But over the longer run the
ability to explore more refined commonalities and their implications is
very favourable to richer linkages in supply-chains (and in more humble
relationships). It also helps refine yet paradoxically thereby simplify
those agreements which so often end up bedevilling partnerships in
old-fashioned EDI. (Think JIT, QOS, and catering for exceptions and
sudden externalities, for example.) (014)
>
> - isPartOf for physical objects vs. for assemblies as designs vs. for
> SysML/UML composite structure and composite aggregation (i.e. Product
> Structure Ontology) (015)
The relationship methods introduced in the 4th instalment provide for
whatever variations you want, though care is needed to keep the lines of
inheritance clear and rationalized. The rel method is one canonical way
to extend the meaning of a rel. (Virtually the entire AOS, by the way,
is composed from canonical rel methods.) (016)
>
> - Reifying relationships vs. Named graphs? (017)
In this canonical world those are merely different views of the same
underlying structure (if I understand you correctly). (018)
>
> - A strong need to for the Dimensions and Units Ontology work to
> proceed (Dimensions are what STEP standards call Property). (019)
Once again as I pointed out in those slides, that is the scene of the RE
or Realword Equivalent, most systematically implemented, highly
modularly, extensibly and most easily reusably. (020)
I can also certainly welcome your Summit presentation's statement of
goals, both its "Near-Term Goal - Harvesting STEP" and the "Future STEP
Project - Semantics Thread". There is so much experience in STEP that
does need to be harvested, and obviously its semantics needs to be
brought up to emerging standards level. (021)
However (of course...) I would prefer to foresee rather a collaborative
project bringing that valuable legacy into the better Agateweb
environment! It's still premature for you to take such a turn now, but
do be prepared to "Ride The Mainstream!" before too long, which I rather
suspect will be before much significant progress in your line with
present technologies. (022)
Anyway, I shall follow your project's progress with great interest, so
good luck with it! (023)
Christopher (024)
_________________________________________________________________
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/OntologySummit2009/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2009
Community Portal: http://ontolog.cim3.net/ (025)
|