xml-dev - RE: local, global (was various ontology, RDF, topic maps)

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

Actually:

1.  Henry Ford was reinventing the production 
assembly line techniques of Olds(?) although 
he is often mistakenly credited as the inventor 
(index of citation noise).  He improved them 
and poured money into them and was said to 
be crazy because of the size of the plants he 
built to support it.  Turned out, crazy like a 
fox.  But remember, he also was the one who 
by dint of the success and investment, had a 
tough time understanding the need to retool 
and change the model.  GM handed him his head 
for that failure.  

2.  Sarnoff invested heavily in AM radio and 
as a result of his manipulations, it took a 
long time (about fifty years) for the superior 
FM format to dominate the broadcast industry.  
Many of you are perhaps too young to remember 
that FM was the broadcast ghetto almost into 
the 1980s in America.

"Worse is better" isn't how things work.  
Simpler does better until knowledge in 
light of implementation becomes understanding 
and gets a sufficiently large community that 
other alternatives become uneconomical 
where an economy is not only money, but 
whatever currency of exchange is recognized 
at the interfaces.   HTML is giving way to 
XML.  Despite Connolly's failure to "grok" 
HyTime, HyTime techniques and concepts 
have emerged and been put to serious use. 
CSS is easier to learn but FOs are gaining 
in currency.  None of these disappeared; 
they awaited a community of users.

It's really an issue of how much how 
many can understand at one time, IOW, a 
thresholding model for a cascade.
 
Ontologies, although expensive to build, 
are already used in knowledge base systems 
(GE used these to make soup) and will become a document 
type required for some kinds of deep 
transactions.   They aren't that easy to 
build and as we have discussed, for widespread 
usage, a document type needs serious authoritative 
credentials and even then, in degree.  An OED 
is needed for some defintions, and Webster 
works for the majority of users.

Them that needs 'em will build 'em. 
Question is, will they share them?  Maybe but 
first the means to share is required and that 
brings us back to Topic Maps and RDF.  Steve 
deRose as I recall noted that xlink databases 
could be a business with a tidy profit margin.

I think the golem of universality is put 
away easily by serious implementors.  We 
just have to be sure the markets understand 
that in terms they can communicate to avoid 
past mistakes and losses due to overhyped 
visions.

Another day at the byteMines really...

Len 
http://www.mp3.com/LenBullard

Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h


-----Original Message-----
From: W. E. Perry [mailto:wperry@fiduciary.com]
Sent: Friday, December 22, 2000 1:23 AM
To: XML DEV
Subject: Re: local, global (was various ontology, RDF, topic maps)


Uche Ogbuji wrote:

> Hmm.  I disagree.  I think Henry Ford effectively scatched the idea that
> there is value in reinventing the wheel, within the scope of his
> operation.  Since then there have been many all-american examples of
> effective, capitalist commerce without wheel re-invention.

Rather than request those examples, let me ask you (and this web-published
list) if
you would feel the same satisfaction with the finality of  Bill Paley's or
David
Sarnoff's 'wheel' as you apparently do with Henry Ford's. Surely the
promulgators of
commercial broadcasting conglomerates would find little value in reinventing
the
wheel which they have perfected. Yet our daily choice of this forum for what
may be
our best or most thoughtful work indicates that we believe there is a place
for a
very distinct alternative. There are also, of course, electric car
evangelists and
mass transit zealots and other stripes of opinion which would take equally
strong
exception to blithely accepting Henry Ford's product (and the ontological
structure
and epistemological viewpoint it embodies) as the final word in its own
domain.

Recall that Martin Bryan said:
Here in Europe we are trying to create a Single Market. The problem is that
this
single market is multilingual and multi-industry. CEN/ISSS has groups
working on
ontologies for engineering, medical supplies, furniture manufacture, shoes,
... There
are significant amounts of overlap in these ontologies, but no knowledge of
what each
other has done or is doing. Trying to get them to stop reinventing the wheel
is a
real problem.

Personally, I can think of no method more likely to produce a new and
unexpected
insight--and out of it, some new best practice--than that multiple
simultaneous
reinvention of the wheel (nor any more certain to suppress a potential
innovation
than blind service to the first principle of a Single Market, or of any
other One
True Way). Indeed Martin Bryan's formulation, above, would be difficult to
improve on
as a statement of the totalitarian pole in the spectrum of opinion regarding
the
proper organization of complex systems. I prefer the foederal approach, and
at the
heart of the foederal approach are statements of the process by which the
foederated
nodes execute specific tasks:  which does what, and in what order. The
detail of
effecting each step is realized at the individual node, and there is
significant
latitude for the idiosyncratic accomplishment of it. The fundamental
mechanism of
adaptability and growth for the system as a whole is that the nodes, without
externally mandated changes at any of them, may be recombined in new
aggregations
with new order of process to accomplish utterly different ends.

So yes, Uche, I unquestionably prefer choice one over choice two. Choice one
responds
to each new problem (indeed, where necessary each new instance of each new
problem)
with the advantages of adaptation at the level of each node as well as at
the level
of the overall order of process, which defines the system as a whole. Choice
two
relies on the authoritative fiat of a canonical solution. Whence derives
this
authority? I dunno. In the case of choice one, the authority is qui fit--it
derives
from the adaptable node within the adaptable system responding successfully
to the
new and unexpected problem as it is encountered.

Respectfully,

Walter Perry