Re: Let's get practical LO601

Jim Michmerhuizen (jamzen@world.std.com)
Wed, 29 Mar 1995 18:55:42 +0001 (EST)

On Wed, 22 Mar 1995, Michael McMaster wrote in LO539:

> Replying to LO417 --
[ ...some material snipped... ]
> My basic assumptions are:
> - the knowledge of an organisation is distributed
> - the knowledge is context specific (including who wants it and who
> is going to produce it)
> - the main knowledge is implicit and can be called forth in
> response more easily than made explicit beforehand
> - it may be unknown combinations of knowledge that are the key
>

I've been working out some analogies between corporate knowledge and W3
logical structures. It's getting exciting. Here's a couple of what look
to me like pretty basic principles:

[1] Abstractly, the original WWW protocol (it's expanded greatly in the
past 18 months) defined just NODES and TAGS, where a node was any
document and a tag was a directed pointer from tagtext to another document.

[2] Even this simple logic is enough to begin representing an emergent
knowledge base, because:

- if nodes are "facts", tags are "meanings". A single fact can
have multiple distinct meanings according to context; and this
is perfectly mirrored in a structure where a single document
can be pointed to from an arbitrary number of contexts, i.e.
other documents.

- there is no prior constraint on what relationships are expressible
within the web structure. Classical relational database methods
depend critically on such constraints -- logically, they are
descended from the formal logic of relations found in Russell
and Whitehead in the early decades of this century.

I hope to be able to put this into practice soon for some people I'm
currently working with.

Regards
jamzen@world.std.com
-----------------------------------------------------^---------------------
. . . . . . . . . . Actions speak louder than words . . . . . . . . . .
. . . . . . . . . . but not as clearly . . . . . . . . . .

>
> Here's what I'm interested in exploring in the area and my ideas about
> approaching it. If anyone knows of anything actually being done that is
> similar, please let me know.
>
[...quote of previous msg trimmed by your host...]