RE: [huml-comment] RE: Human Markup Language 1.0 considered harmful

["Dennis E. Hamilton" <dennis.hamilton@acm.org>]

<ORCNOTE> Inserted below. </ORCNOTE>

-- Dennis

Dennis E. Hamilton
AIIM DMware Technical Coordinator
------------------
mailto:dennis.hamilton@acm.org  tel. +1-206-932-6970
http://DMware.info/             cel. +1-206-779-9430
     ODMA Support http://ODMA.info/


-----Original Message-----
From: Rex Brooks [mailto:rexb@starbourne.com]
Sent: Wednesday, November 27, 2002 05:54
To: rkthunga@humanmarkup.org; dennis.hamilton@acm.org;
humanmarkup-comment@lists.oasis-open.org
Cc: 'William Anderson'
Subject: [huml-comment] RE: Human Markup Language 1.0 considered harmful


[ ... ] our
specification is not intended to specify definitions for terms in
common usage, but to incorporate them in such a way as to be
interoperable.

<ORCNOTE> An example of how you think that is accomplished would be very
useful.  In my experience, removing the contextualized use of something like
the attribution of race (or gender) is not going to provide
interoperability.  It just means stuff will be used out of context and the
collisions will be harder to detect.  I don't see how that is an
improvement.  I'd like to see how you will avoid that.  </ORCNOTE>

[ ... ] We took a year and a half of very intensive work to
produce the smallest possible set of base terms which serve mostly as
placeholders for essential concepts from which more specific terms
can be derived and put into applications in a way that allows for
some standard descriptions of human characteristics and human
communication.

<ORCNOTE> Lemme see here.  There are base terms.   That's what we have now.
It is thought to be the smallest possible set.  These are placeholders for
essential concepts.  Uh, do we have the concepts yet, or just placeholders?
(Hmm, terms without concepts.  Never mind that.)  OK, and then more specific
terms (concepts?) will be derived.  And when they are put into applications,
there will be some (or at least the prospect of some) standard descriptions
of human characteristics and human communication. How am I doing?</ORCNOTE>

<ORCNOTE> It is hard to see how refinement is going to give rigor and
precision if the top level isn't rigorous. </ORCMID>

<ORCNOTE> Whatever framework is set at the top level, I assume that it will
guide and limit the kind of conceptualizations that can be achieved by
refinement and specialization, if that is what is meant here.  I believe
that is what happens with terminological frameworks.  So I am nervous when
we talk about how it will get worked out in the application. </ORCMID>

<ORCNOTE>
If I have any insight to offer here, I think what is needed is the
metamodel.  That is how are variety and diversity and all of that unruly
stuff to be dealt with so that collisions of different uses of seemingly the
same terms are recognized.  Or is this not a matter of concern? </ORCNOTE>


As for self assertion, it stems from a problem which we avoid
entirely, authentication, authorization and certification, and the
prospect of testing to see if there is any way to determine whether
any given entity is a currently living biological human being
actually engaged interactively with a computer at one end-point of a
real-time communication interaction on the Internet or in any given
network. When we simply accept self-assertion and ONLY begin to apply
our language in applications AFTER some authentication has taken
place, if it has taken place, and whether or not it has taken place,
we make it possible to simply evaluate the online behavior (if that
is part of an application using HumanML) against standards that are
as much as we can possibly make them, unbiased  culturally or in any
other sense we can reasonably avoid.

By doing that, we make it possible to achieve a greater degree of
accuracy in evaluating communications and, ultimately, in formulating
communications.

<ORCNOTE>
Granted that you're not developing a specification about authentication,
authorization and certification, I don't see how you can claim a greater
degree of accuracy (I am not even sure I understand what it means to say
accuracy here, but I'll play along) absent some representation of the
authority and provenance of an assertion.  How can one evaluate a
communication without that?  Back to the metamodel question.  What is the
model for incorporating or reflecting context/situation in some federating
way that allows your "smallest possible set of base terms", which is to say
"smallest possible set of base concepts" to be developed as orthogonal but
not independent.  (Wow, I made a trope!)
</ORCNOTE>

[ ... ]

Ciao,
Rex

<ORCNOTE>Ci vediamo.</ORCNOTE>

P.S. We have a spot warm for you if you wish to have your opinions
count a bit more. But public comment is appreciated, so thanks.

<ORCNOTE>Prego.  I'm playing hooky as it is.  Thanks for the invitation. I
decline for now. </ORCNOTE>

[ ... ]