RE: Taxonomies, URN's etc..

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

Just do a basic system of physical descriptors 
and stop.  That's one module.  It can be 
extended later.  I am ready to be convinced otherwise but 
only by coded examples.  Do not constrain the 
use, do not become the repository, do not otherwise 
open the scope to initiatives which we have neither 
the resources nor commitments to manage.

What I coded was an example used between a 
fingerprint system and an arrest database.  It is 
real, it is working, and it represents a customer 
use view of data.  It was provided only to show 
that identification is a process, not a data type 
that is independent of a system.  What we need now 
are examples just as the EMOTE data provides a 
working example of a set of parameters for communicating 
with a character animation. That enables us to 
look at our prototype schema fragments and determine 
where extensions might be made.   A solid data description  
will be far more reusable than a complex class 
description wanting implementation.

We are already tied to OASIS.  Leave that exactly as 
it is and do not open more liaisons unless the 
community approaches this one to work under the 
OASIS authority.  Otherwise, we are starting a 
never-ending project that will not deliver.

Len 
http://www.mp3.com/LenBullard

Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h


-----Original Message-----
From: Rex Brooks [mailto:rexb@starbourne.com]

HumanML will probably first 
tackle the area of Human Identifiers, including physical 
parameters--for personal medical information--and certification, 
whether as a certifier itself, heaven forbid, or as the arbiter of 
such through the auspices of the only totally non-profit, 
non-proprietary, uncentralized, and also, non-governmental and 
therefore unbiased, in legal terminology, repository of information 
authorized by individuals themselves.

To understand that think of Profiling. Whether creating a profile for 
a Human Object that is the representation of a software agent or an 
actual person, HumanML will have a breadth and depth that allows for 
more areas of information than other database-centric systems created 
primarily for authentication of identity purposes such as Len 
described yesterday. As such HumanML will likely play a role. What 
that role will be is far from clear, but that is certainly an area 
that will deserve some dedicated work.

One question which came up was longevity, and I suspect that we need 
to establish ties to as many academic and governmental computer 
installations as possible in order to ensure that our DISTRIBUTED 
repository structure will have a life of its own as long as the 
Internet does at least. Tying ourselves to OASIS or any specific 
industry consortium or group or association, is fraught with the 
typical dangers of business in general and the politics of consortia 
in particular. They are actually fairly stable, but while OASIS is a 
bit of an exception, they tend to be braking mechanisms, serving to 
prolong the shelf life of extant business product.