Lists Home |
Date Index |
From: Rick Jelliffe [mailto:firstname.lastname@example.org]
>I think James is setting the stage for the ISO Document Schema Definition Language effort.
Ok. I think ISO has the obligation to harmonize as it goes. Just as I and others have
critiqued the W3C for having hodge podge specs and it has replied with the TAG, ISO
should consider ensuring that this effort and a modernization of ISO 8879 occur together.
I realize the enormity of that, but it would be a giant leap forward. I would think that
if ISO 10744 is to be relevant in some way, it's editor and supporters would also
participate. If as some (James Clark) suggest, that is a rather small group of
people, the coordination shouldn't be difficult and perhaps more prudent than a
"stab in the back".
>This is intended to be a simple but general framework for SGML, XML and their friends,
>neighbours and successors for doing things like
> * associating documents and DSDL processes
> * listing sequences of validating and transforming (V&T) processes
> * standardizing some well-known V&T (Schematron, RELAX NG, W3C XML Schemas, etc.)
> * creating the slots into which new V&T can easily be added (e.g. a refactoring of
> W3C XML Schemas to make it more modular.)
I like the look of that. Please expand these acronyms: DSDL, V&T.
>It builds on various successes and failures in this regard:
> * DTDs and attribute defaulting was successful
> * SGML's "LINKS" (a method to switch in and out banks of attribute declarations
> based on various criteria) was not
> * SGML's Archictural forms was not
Architectural forms have been successfully implemented. See Dr. Newcomb's earlier
posting. Even if not adopted as part of the ISO effort cited above, the requirements
and results of the implementation should be reviewed and a cogent summary made. The
specifications for this originate in ISO. Again, ISO should be held to the same
responsibilities being charged to the W3C. In fact, as an international standards
organization, it should be even more conscientious.
>I think Archictural Forms failed to thrive for several reasons:
> 1) it went against the dominant need of the mid-90s, which was for
> minimal SGML to get smaller, not larger.
I accept that. Timing is a killer. Yet here we are reexpanding the
horizons after making claims about how simpler was better.
> 2) it was put out as part of HyTime, and so its message was diluted or buried
HyTime grew like topsy but also stayed in committee a very long time. Just as
we ask for the W3C to refactor, it should be refactored if it is to remain even
if that means submergence in a new standard. We have all at one time or another
had to surrender our work into other efforts, but where that surrender comes
at the cost of public or private humiliation, we only set the stage for another
conflict. Humans work on these things. Respect is a greatly valued quality.
> 3) it is still grammar- or tree-based, and so did not adequately address
> the biggest lack of DTDs: non-regular constraints
> (just as SGML's LINK feature was inadequately powerful to be much use)
The co-occurrence constraint problem was not posed as a requirement to be
solved, in my memory. Yes it should have been. That will be addressed
in the ISO committee you are part of, and I suspect, solved.
> 4) as species of transformation language, it is not powerful enough.
It may be that too many tasks were addressed. That is certainly something we
are now all familiar with, but in those days, the SGML experts were a very
small group of people trying to do a very large task. Yet since many of
the same people are working this problem, I can only hope experience has
engendered clearer thinking on these subjects.
> 5) it invented its own non-SGML dialect with a meta-DTD containing
> <!ADFR ...>, a new particle #ALL,
> and new parameters on the SGML Declaration.
It seems they were trying to solve this all within the context of making
as few changes to ISO 8879 as possible in the same way the W3C tries to
layer an infrastructure for XML without touching XML 1.0. Again, I think
it wise to at least consider a reworking of ISO 8879 at the same time if
that is necessary. Since this is all occurring under the aegis of ISO,
this should be doable.
>However, architectural forms have not died: in fact they live
>on through W3C XML Schemas. It allows you to specify
>an abstract grammar in which gaps are allowed in content models
>for elements (e.g. from other namespaces) and which
>(through subsitution groups, etc) you can specify that the particular
>names in a document are aliases for some other names (or
>types), and which lets you augment the information set with
>other attribute values.
>Even Schematron has a role attribute, which could be used to augment
>an information set for the benefit of downstream processes. I think
>the path approach is easier to figure out and more powerful than the
>meta-grammer approach of AFs.
>So I certainly buy that the AF approach of a schema language for
>expressing patterns that underly the explicit markup is good.
>However, I think now that we have a similar thing expressable
>using XML Schemas (in a form apparantly amenable to OO
>programmers) and with Schematron (in a form amenable to
>most people AFAICT), I cannot imagine that AF per se has much of
>a life. Archectural Forms has lost the battle but has won the war.
Then there is baby in the bathwater in the form of the requirements
to solve certain problems. If we can duck the knife fight that will
start over "who invented what" that XML brought on itself and which
continues in some dark alleys of XML journalism, there is a chance
that all of the good work done for that last two decades in markup
schematization can converge in this ISO effort. I earnestly plead
as I have before with the ISO members to work together toward that
end confident that what they have contributed to the community will
be honored, respected, then stripped, retooled, and assembled into
a sleeker and modern system.
One can hope there is a better way than the blade, but it comes
down to individuals.