Pete Johnson <[log in to unmask]> while talking to
<[log in to unmask]> and <[log in to unmask]> and others says:

[snip lots]

>My area of interest in this is to aid decision support systems - this
>means using the EMR for a task which the user did not have in mind (most
>likely) when he recorded the entry.  The reason why I still pursue this
>goal, is that there are workers (only a few, I admit) in AI who believe
>that it is possible to get to the holy grail of task independent
>knowledge, and a great many who believe it is possible to get close to
>this, if not actually reach it.  What we should be able to achieve is a
>great improvement on the current situation.  That may be enough to achieve
>what we want.  But I'm sure the use of SGML per se isn't going to solve
>these problems.

This is an interesting set of remarks. SGML doesn't solve anything on its
own, because it is an enabling, but non-directional technology, which sets
out to make knowledge tasks independent. How well it does this is a matter
of insightful, and perhaps still virtuoso, use of tagging and the DTD
convention.  As with programming in general, there are at least 3 levels:
adequate (x1), good (x10) and off-the-wall expert (x100).  Each level is
at least an order of magnitude better (against any measure) than the
previous one in what can be done with a given tool.

Tagging is a classic AI means of enriching data and information to create a
knowledge base. SGML offers that potential in a particularly flexible way.
Task independence is only 1/2 of the coin: intelligent agents are then
needed to harvest information for a particular set of problem-solving
issues. Various tools have been developed with this at least partially in
mind: viewers, DSSSL, HyTime, architectural forms, maps, etc. Each, of
course only embodies a partial generality and so, together with SGML,
they only piece-wise (?perhaps asymtotically) "approach" full data
independence (if such independence requires a single, all encompassing
framework and set of rules).

The advantage of SGML is the tagging open-endedness that can be applied to
any document as an information container, either by modification (where
allowed), or by overlays.  Moreover, SGML annotated documents, as others
have noted, fit well as the fundamental "units" within an information
management structure based on familiar relational and object-relational
principles where they can be handled in conventional ways.

The other advantage is that much of the potential sophistication of this
approach is in the markup background. Thus reports of this sort can be
treated as if they were the visible text only (or with only certain tags
operative as format or field markers).  Thus the bottom-up introduction of
SGML does not require a "FEDERAL EXPRESS" type of configuration where
nothing can be gained until the whole system is in place.  Rather it is
ideally designed for the incremental "colonization" of the present
generation of legacy healthcare systems. Evolution rather than revolution.

Tom

 p q
 \|/
 /|\   TOM LINCOLN  [log in to unmask]
 \|/  "Life is short, art long, opportunity fugitive,
 /|\   experimenting perilous, reasoning difficult."



%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%