On Tue, 13 May 2003, Peter P. Jones wrote: (01)
> > I wouldn't call it fogging the picture, I'd call it, perhaps,
> > layering.
>
> Which implies that distinguishability is retrievable. That's fine as
> long as the algorithm for retrieval is not hidden. (02)
I suppose that depends on what you mean by algorithm. (03)
Is there an algorithm for human communication? I don't think so.
I think people talk and stuff happens, sometimes boring,
sometimes interesting. (04)
Wiki is a _tool_ not a method. It can be a tool that is a part of
a method. Like most tools it works for some people and not for
others. There is no one true tool. It's clear that it doesn't
work so well for the goals you have mentioned (natural language
processing, well defined classes of information rather than fuzzy
categories of context). (05)
A tool like Wiki augments certain styles of behavior. For me
these are note-taking and connecting ideas. (06)
I like this because I was (happily) brainwashed by a physics
teacher in high school who said that the true genius is someone
who can draw connections between seemingly disparate ideas. (07)
> > WikiWords are concepts are names are patterns are containers are
> > objects are neurons are transmitters are identifiers are labels
> > are categories ...
>
> But that's precisely the kind of overloading I'm fighting against. (08)
Yes, I understand that. With your goals that overloading leads to
dissonance. With my goals that overloading leads to harmonic
resonances. (09)
What's great about all this, though, is that I can go around
creating piles and shouting eureka every now and again and you
can eventually gain something from my product. Meanwhile, you can
create clear, well-defined ideas that help other people like you
as well as people like me who will stack them up and squeal with
delight and say, "See, look, it's really AllTheSameThing". (010)
And thus we engage in communication that leads to the creation of
knowledge. It's messy, sure, and it can be improved, but it is
how these things work. (011)
Tools like Wiki, Blogs, PurpleNumbers improve things by creating
referential networks of exposure that I think inspire lateral thinking. (012)
Tools like RDF, NLP and other "rigorous" friends alongside
traditional classification and IR improve things by facilitating
retrieval; question answering. (013)
This is a classic dichotomy in library and information science.
It makes for great arguments because it tears right down to the
core of what we are trying to do. It's encapsulated in the
question: What's a better library: one where you can browse the
shelves, or one where you use the card catalog to find a book
which you then must ask someone to retrieve from the stacks. (014)
I like the former, but I understand that both are needed. (015)
My personal philosophy, though, is that when trying to solve
wicked problems that are so wicked they don't really have a name
one puts a significant limitation on oneself if one chooses to be
in a world of conceptual clarity. We don't need answers to
questions we are able to articulate. That is we don't need to do
information retrieval. Rather we need to create ideas based
analogy, connection, comparison, contrast to be able to start
asking questions. (016)
Once you know the questions, boom, it then becomes necessary to
spend some time in the land of conceptual clarity: questions need
to be asked and answered, hypotheses need to be tested,
information needs to be shared with clarity... (017)
As far as I've been led to believe, those two paragraphs describe
innovative science in action. (018)
Anyway, Wiki can be a tool in this process, for some people. (019)
> > Once you have identified a stack of things in a layering, you can try
> > to apply the associates of an element in the stack to one of the other
> > elements in the stack. This leads to questions and answers that one
> > may not have thought about before[1]
>
> This point is interesting... Can you provide an example? (020)
That's what these are: (021)
> > Okay, so say a WikiWord is a neuron. What might be LSD for Wikis?
> >
> > A WikiWord is a label. Human readable labels lead to namespace
> > collisions. What can we do about that? Is it a good or bad
> > thing? [2] (022)
The first one was mostly a joke, but a lot of good thinking comes
from jokes. (023)
Maybe we can get Jack to step in on this one: He and I have both
worked on Knownspace. In the original data model there are things
called Entities. Entities have Attributes (which are also
Entities). Jack has in his head a model (or maybe he did it?)
that uses tuplespace for the Knownspace data model. Making that
change changes the action model in the space. How it changes
exposes things about both models and provides insights into areas
of improvement. (024)
Another example: (025)
A SoftwareDesignPattern is like a ShakespeareanSonnet. I've used
this comparison to help explain what a DesignPattern is. (026)
To take that a little further: review of design patterns can show
they are, from several thousand feet, really all the same thing.
So it might be more appropriate to say that a DesignPattern is a
Sonnet and Command and Composite are Shakespearean and
Petrarchian. (027)
An insight from that, then, might be to inquire what a limerick
design pattern might be. (028)
> > If you want to be more formal about what I'm saying, change 'are'
> > above to like-a rather than is-a and then place it all in the context
> > of my belief that productive thinking comes about through analogy.
>
> That would be better from the point of view I am maintaining. But
> then I would have to ask why that move is not explicit in a humble
> mashed Wiki word? (029)
The mashed wiki word itself does not provide the analogy, it is
the network of connection they help create. The mashed words are
just a tool that greases the creation of the network (for some
people) by providing external cognitive marker spots. (030)
Uh. Hmmm. I have no summary. (031)
--
Chris Dent
cdent@blueoxen.org (032)
--
This message is archived at: (033)
http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=tools-yak&i=Pine.LNX.4.44.0305141257320.21434-100000@cream.burningchrome.com (034)
|