Group Admin: Please remove my email from this list.
On Fri, Dec 7, 2012 at 9:08 PM, John F Sowa <sowa@xxxxxxxxxxx> wrote:
On 12/7/2012 10:39 PM, Hans Polzer wrote:
> My point was that it's not just a matter of evaluation criteria.
> It's also the definition of the context and associated scope in
> which the criterion might be applied, as well as the context
> and associated scope in which the ontology might be applied.
I agree. I was responding to Steve, who said "ontology *evaluation*
is a more focused topic".
I also agree with all the other points you raised:
> Looking at both the assessment context and the contexts in which the
> ontology in question is to be applied will provide a basis for deciding
> which of the potential assessment/evaluation criteria, such as the types you
> list below, are most appropriate in a given assessment context. "What
> difference will it make" is a good question to ask in order to determine
> whether a specific criterion makes sense for a given assessment or
> application context attribute value range. Some criteria might be
> universally applicable, i.e., a "core" set, while others will be appropriate
> in specific assessment contexts and in specific application domain contexts.
At an early stage in my career at IBM, I was involved with evaluating
various competing computer architectures. Some were clearly worse
than others for many reasons. But there was no linear scale of
goodness for any of them.
As time went on, the reasons why one or another became more widely
used had nothing to do with how good they were.
As just one example, nearly everybody -- including Intel -- realized
that the X86 architecture was a disaster. Even today, it still
inherits a pile of crud that evolved from the 4-bit 4004, to the
8-bit 8008 and 8080.
Intel tried to kill the X86 *twice*. In the early '80s, they tried
to replace the X86 with a better 32-bit design, but the popularity
of the IBM PC made it impossible for Intel to kill the X86.
Then in the 1990s, Intel wanted to replace the X86 with the Itanium.
But AMD almost beat Intel with the 64-bit extension to the X86.
That forced Intel to keep pushing the old X86.
Digital had a real winner with their Alpha chip, but their minicomputer
business was killed by the PCs, while they couldn't break into IBM's
mainframe business at the high end. Then Compaq bought Digital, but
they had no idea what to do with the Alpha chip. Then HP bought
Compaq, but HP management pinned their hopes on the doomed Itanium.
If Compaq or HP had sold the Alpha chip to a company that knew what
to do with it, they might have made it into a winner. But it died.
Morals of this story: (a) evaluation is extremely difficult,
(b) there is almost never a linear order of goodness for anything,
and (c) when you have a potential winner, it can very easily lose
for many reasons that are irrelevant to its inherent goodness.
--
Stephen Ufford
Co-founder, CEO
Trulioo
Direct: (415)845-0730
eFax: (714)551-2588
www.trulioo.com
_________________________________________________________________
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/OntologySummit2013/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2013
Community Portal: http://ontolog.cim3.net/wiki/ (01)
|