KnownSpace gets so many things so much closer to right than anything else
I've seen. (01)
For instance, when looking for ways to capture bits of information in ways
that rapid mapping to any variety of so-called knowledge representation
schemes (e.g. DAML-OIL, topic maps, RDF, frames, whatever), it appears that
the KnownSpace entity-attribute mechanism is about as generic as anything I
have been able to imagine. (02)
The built-in extensible agent architecture satisfies nearly every
requirement I have been able to imagine. (03)
What has been left, imho, has been the nature of that architecture in light
of a distributed system, one in which there is seamless federation of
installations. The KS crew is developing their event architecture in a
direction that may well enable such federation. My experiments have been to
substitute tuplespace for events. I'm far from being able to make any
claims, but I suspect that some sort of tournament between event and
tuplespace architectures is in order. For now, though, the event model is
rapidly maturing. (04)
Cheers
Jack (05)
At 07:15 PM 12/19/2002 -0500, Chris Dent wrote:
>You may have seen Jack's announcement about the
>Helium/Knownspace project I'm working on. Now that we are at the
>end of the semester several of my compatriots and I have realized
>that while the Knownspace architecture is awesome, the
>overwhelming weight of the life-changing hype associated with it
>actually detracts from discovering concrete, step-wise
>improvements in life that could be gained by using the
>architecture. (06)
---------------------------------------------------------------------------
XML Topic Maps: Creating and Using Topic Maps for the Web.
Addison-Wesley, ISBN 0-201-74960-2. (07)
http://www.nexist.org/wiki/User0Blog (08)
--
This message is archived at: (09)
http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=tools-yak&i=5.1.1.6.0.20021219162448.0240b8f0@thinkalong.com (010)
|