[xtm-wg] RE: [topicmapmail] PSIs of general communication

["Paul Prueitt" <bcngroup@erols.com>]

Well,  do not know what Opera is, and have not downloaded Netscape.  I will
allow this to stop me from looking at the site.

These browser wars are really an inconvience.

I am just exploring and do not know what I am looking for.  The XTM 1.0
specifications is helping me a little.

The issue I think I am most interested in is semistructured data where the
strcuture have the complexity of natural langauge grammars.  I would like to
have a query language into natural text.  (smile).

I forward some thoughts on this, and look forward to additional discussion
at the Auston conference (March 4 - 7).

****

Consider the kinds of information searches that Jack Park refered to (in
kmci-Virtual-Chapter forum):

"
1)  Perhaps it will let me search for which manufacturer makes a vehicle
that
 will carry more than 130 passengers for more than 1500 miles for less than
 $0.7/seatmile.

2)  Perhaps having told me who the innovators were it can tell me what were
 their learning styles and what they did to transfer their findings to
 application action agents and how long that took.
"

are exactly the type of search that NCorp and IJEN are claiming to have.

www.ncorp.com   and www.aboutijen.com

as I understand it, Michael Lynch (Founder of Autonomy) is caretaking this
new company - trying to integrate the Autonomy Shannon + Bayesain + neural
network information representation with some serious e-commerice and B-2-B
things.

The idea is that by moving the realtional database data into a vector space
that then operating in that space, then the notion of nearness is
"fuzzified" in a way to allow such serach and retrievals.

I think that the concept is good, and there will be some value.  However,
for me fuzzification is a poor substitute for sub-structural affordance
analsysis, and NCorp keeps the relational database as the "addressable
information resources" (to use the terminology of the XTM 1.0
specification.)

http://www.topicmaps.org/xtm/1.0/index.html

I know of several separated technology innovation processes that would
replace the Relational Database completely.  This is very exciting stuff.

Denham, I feel that you uniquely could define a set of "queries" that are
not addressable in SQL, are better addressable in IJEN query and are even
better addressable in topic map query language (TMQL) [which we must now
invent.].

The question that I have talked about with various community (TIPSTER, TREC,
CIA, etc.) is "which of my three daughters do I love the most?".  The answer
can not be in the form of a ranked list even if the context is perfectly
expressed.  Thus the answer is just the context only.  I feel that the
answer can be answered using my tri-level machine architecture and the
voting procedure because there is the use of rough sets to define
overlapping and nested category boundaries.  (Like inner and outer measures
on sets of real numbers - but in category theory).  The overlap creates
structural affordances that cause secondary placement (similarity)
criterion.

"What is the largest integer" and "what is the smallest position real
number" are two other interesting queries.  In these cases, the query can
lead into an investigation of the meaning of the notion of nearness and next
to, and thus produce (as was the historical fact) modern mathematical
analysis and much of topology.


------------------------ Yahoo! Groups Sponsor ---------------------~-~>
eGroups is now Yahoo! Groups
Click here for more details
http://click.egroups.com/1/11231/0/_/337252/_/982589728/
---------------------------------------------------------------------_->

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com