[self-interest] programming and knowledge representation

[thilo.s at web.de]

Brian Rice wrote:
> 
> Definitely so, although arrow is on the "back-burner" while I make
> sure Slate works properly. Slate embodies some arrow concepts, but
> not even close to all of them. Arrow is an information system that's
> not really classifiable by the current state-of-the-art in
> information (or "knowledge") systems research. Unfortunately, my
> thesis papers on this subject are far out of date compared to my
> current state of knowledge on this (which has grown by orders of
> magnitude). 

Do you mean your grown knowledge on Arrows or on the knowledge systems
research ?
Anyway - don't you think there could be a common ground for us to
discuss about ?
I feel your ideas are more advanced then mine - but (of course this must
be harder to acchieve) not so clear (to me).

> However, I don't believe that this discussion is really appropriate
> to this list. 

Yes. Partly. The problem is IMHO that we are exchanging knowledge via a
medium that forces us to (statically) categorize - something that
selfers don't like to much anyway ? ;-)
Categorizing by the means of having a "motto" (here self-interest) for a
group - is seducing us to measure a kind of distance (what is the metric
anyway?) of contributions to the motto. The motto is IMHO just a very
primitive way of giving us the opportunity to filter
knowledge/information - in advance and applied by the author (in this
case I set this filter because I think that people could help me a lot
advancing my knowledge in especially this group - kind of selfish
thinking - but hopefully there will be a gain also the other way round
:-/ .

This knowledge-categorizing is the problem - I could start yet another
egroup or I could move to comp.ai.philosophy or/and another group - the
problem remains unsolved (I already thought of posting everywhere - also
resulting in getting more people to know this egroup :-).

Anyway, the next step in filtering is setting up rules in your
message-reader system. Do you really use it ? I am not. Because the
system does'nt let me apply the same filters for different groups -
because email-filtering works differently then usenet-filtering etc...
Of couse I could switch to a more sophisticated system - but still there
would be other disadvantages - because we still rely to much on text and
categorizing.

So this is a problem of usenet in general - and moreover of written,
categorized text.
This is exactly what future systems should get over with - and what my
(and I think Tunes) ideas are also about.

> I'd rather hear less about the vague notions of
> "knowledge engineering" and more about language and system design
> issues. 

Yes, in text it gets all mixed up - but only because things are really
interwoven (IMHO).

> All knowledge is expressed in some language, whether at one
> level or another. A language expression might evaluate into a
> structure which is the information, or the expression itself might
> include or encapsulate the information in various ways. The
> intricacies of how this is done is embodied by studying what "meta"
> means in context, which incidentally is always relative to some core
> calculus, which in Self's case is not the language itself. The Self
> language in fact consists of several distinct phases of speaking
> about objects, which is part of my interest in it.

Could you explain what you mean by phases here ? Like views/aspects on
objects or kinds of objects?

Regards,
Thilo.