XML Information Set Requirements, W3C Note 18-February-1999

[len bullard]

Rick Jelliffe wrote:
> 
> Do you mean "documents" (packages of structured information) or
> "electronic publishing"? I think you mean the latter.

I think they are hung up in the venerable "what is a document" 
argument.  The answer is, whatever you want it to be as long as 
you stick to the markup standard.  That's not hard.

> XML-like SGML has been used for years in applications other than
> publishing: someone told me that Xerox have used an XML-like syntax to
> deliver copier diagnostics to repairmens' PDAs for almost 10 years now.
> (Can anyone confirm this?)

I can confirm they used an SGML-like tagging language for a print
system.  It was famous 
for excluding attributes in the design.  It was deployed at USA MICOM.  
It was considered "that thing in the corner" because in not sticking 
to the standard it presented it's owners with island of automation
problems. 
Markup was adopted to integrate documentation production and
distribution processes 
and to enable information lifecycles that were longer than 
the lifecycles of the host systems.  There has been over the years 
many attempts to adopt a procedural/programming design in the 
context of markup.  That is another long story.
 
> HyTime was encouraged in part because of CIA interest in languages for
> orchestrating satellite movements, I have been told, too. Indeed, HyTime
> grew out of a desire to formally analyse performances of music.

That is close.  HyTime had its origins in the desire to create a music 
description language.  That necessitated a timing model.  It was
postulated 
by several interested observers (from CIA, CALS vendors, and a fellow 
from God's Brain (inside joke)) that a generalized timing model which
included 
synchronization could be applied to managing very large and distributed 
enterprise processes which included NASA launches, process/control 
design for manufacturing (simulation), and so on.  Indeed, for the
general case, 
this was true.  

HyTime had several areas of interest including defining the general
classes of 
hyperlinking and addressing.    Timing models for real time systems are 
difficult to generalize because of issues like "continuous vs discrete", 
what to do about race conditions, event fanout, etc.  The VRML community 
took these same issues up.  It is very thorny across a network with 
unpredictable delivery (eg, the WWW) for distributed simulations.  To 
summarize, the HyTime models could probably be used for documenting 
historical performances, but might not work well for real-time control.
 
> Music performances, satellite movements, diagnostic data: these are not
> "publications" but they all can be "documents".

It may be that the reverse is the case.   Given what markup does best, 
it may be that it is indeed a publication.  That begs the question of 
"what is a document" but that question never gets answered outside the 
parameters of what the standard defines.  The point at which you take 
the M out of XML, you are defining a different metalanguage for a
different 
set of requirements, therefore, a different standard.  This is not the 
charter of the XML Information Set WG.  They have a constained task  
task which the HyTime efforts have proven is hard enough without more 
rabbit trails.

Think long and hard.  There are several interdependent efforts moving 
in parallel and other language efforts outside the core XML standards 
which depend on these (eg, X3D).   If rabbit trails impact these
deliverables, then 
the decisions by the W3C and W3D consortia to close the WGs to members 
and invited experts are justified.

len

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)