Re: [ba-ohs-talk] OHS/DKR Design for KM and Licensing
Rod, (01)
The web site cited below is not part of any particular project. Rather, it
was created as a kind of hyperlinked space where I could collect my
thoughts using the editor that comes with the Mozilla browser. Now,
however, I have a new system that is awfully close to finding its way to
the web, and that system is essentially a Java Wiki on steroids. I hope to
have much more to say about it soon. (02)
I ought to say something about the relationships between various thought
streams and the visible project Nexist. Nexist is about to get its own
home page at http://www.nexist.org (don't go there any time soon, the
server is sitting on the desk next to me right now). Nexist was, is, and
will always be an engineering prototype with which various strategies for
wiring up communicating agents can be developed. The goal there is to find
an API that allows this to happen; I have said here many times in the past
that what a proper OHS needs is some sort of API (or maybe collection of
them) such that developers have something to develop against as they add
value to the project. I'm not quite sure on this but I think that
r-objects.com either has such an API or is close to one for Pepper. I
think that to be a valuable notion for the onset of any OHS evolutionary
process. (03)
My more recent projects set the agent technology aside in favor, but not
abandoned to, a search for user interface technology. I've recently come
to the idea that exploring how far one can push Wiki technology seems a
worthy pursuit. In the end, I think this work will then get married to the
Nexist API (or whatever evolves along those lines) and the Wiki project
will become one of the many agents in a mix that couples users to
cooperative knowledge building activities and infrastructures, as would be
an OHS. (04)
I should also point out that Nexist, the project, is about to go prime time
as my book _XML Topic Maps_ is now at the printers. I've seen it at Amazon
and Barnes & Noble. Nexist is one of four open source topic map-related
projects discussed in the book. (05)
If I might expand on the quest, let me say that I think that a proper OHS
project (make that projects) will involve creativity and evolution on
several fronts, including:
user interface (frontside)
knowledge representation
information transport
knowledge accretion (middleware)
search, mining, concept formation, etc
back side support
persistence, reliability, security, authentication, etc (06)
In my view, information transport, which couples all agents operating in an
OHS environment, is the proper place to be looking for publishable APIs. (07)
I do not think user interface is going to be easy. First, that's because
no two individuals on the planet are liable to necessarily be happy with
precisely the same user experience. Doug's call for user control of view is
of great importance here, and I would add navigation and other factors into
that. Second, that's because the range of use cases an OHS might need to
support could mount to something enormous. Consider the idea that, as Ted
Nelson has always said, there should be no difference between the browsing
and the editing experiences. Early Macintosh junkees wore t-shirts that
said "Don't Mode Me In." That one's a devil, particularly if you are using
HTTP protocols, as my Wiki presently does. Look closely at r-objects
Pepper and you may find evidence that this particular problem is on its way
to being solved. There really is hope out there for useable interface
concepts that can be adapted to an OHS. And, then there is the issue of
just how users want to approach the Web. There's blogs, outliners, Wikis
of all kinds, forums, email forums, and so forth. Strikes me that you
essentially must satisfy all those needs if you're going to get sufficient
buy in to OHS technology. (08)
From my perspective, I will not be happy with anything less than a
full-monty emanuensis, one that can help me to keep track of all of my
hair-brained ideas, do Web searches on key words and phrases it finds that
I use, and can take parts of my personal knowledge base to the Web to synch
up with others, either through a Portal, or by way of P2P. The KnownSpace
project http://knownspace.sf.net
is a great experiment in cooperative agents, security and a variety of
other issues. Murray Altheim is making progress on a tool that will do a
chunk of what I want, perhaps a big chunk. Lee Iverson is progressing on
his backside system NODAL. That's going to help. Toss my Wiki into that
mix and I tend to think I'm coming awfully close to having the tools that I
want. Come to think about it, SDS represents concrete steps in the same
direction. (09)
While on the notion of my perspective, let me point out that there are no
doubt several (dozens?) of individuals reading/lurking/contributing on this
list that share with me some visitation with a life-threatening disease. In
my case, the Leukemia is in really deep remission. I beat it by applying
massive intellectual horsepower to the problem (all my physicians, etc),
biotechnology, and so forth. But, realistically thinking, it's not beat.
It's sitting there now evolving, maybe into something worse. Others on this
list no doubt face the same "clonal evolution" as well. My strategy: get
started using an emanuensis, team up with others, and try like hell to stay
ahead of the game. In almost every sense of the words, I think that is
what Doug is calling for on a much larger scale. (010)
Cheers
Jack (011)
At 07:01 PM 6/5/2002 -0700, you wrote:
>Jack,
>
>Thanks for sharing progress on your system for a collective learning
>machine (CLM), cited in your letter on May 17. As you add content to
>links on your web site, please share examples of work product showing
>how learning is improved through the collective model being
>developed. Is this part of Nexist or a separate project?
>
>Thanks.
>
>Rod
>
>*************
>
>Jack Park wrote:
> >
> > A while back, I began to collect my thoughts. Far from complete, and, in
> > fact, I haven't touched those pages for a while now. However, perhaps it's
> > time to reveal them and see what might come of that.
> >
> > http://www.thinkalong.com/JP/cpc/
> >
> > Cheers
> > Jack (012)