Hi Don -
I'm assuming that "subsetting the DTD" refers to the proposal to provide
a simplified CD 1.1 DTD or CD 1.1 "lite."
You probably are aware that the Todd Vincent who expressed the
"over-inclusive but optional" approach to developing the Court Filing 1.1 DTD
is the same Todd Vincent who, after working with the CF 1.1 DTD in the GCAC
interoperability pilot, wrote the "Lessons Learned" document. That paper
suggests that "It may, therefore, make sense for individual jurisdictions to
create a 'backwards-compatible subset' of the Court Filing DTD for specific
purposes and requirements."
For clarification, is your concern that "subsetting" the CD 1.1 DTD
should not be addressed in the CD 1.1 spec itself but should be left to
individual jurisdictions to address via court policy standards? It sounds like
the worry is that subsetting the CD 1.1 DTD as part of the Court Document
specification may intrude upon and get crosswise with court policy
standards.
Or is your concern broader - that there should never be subsetting of the
CD 1.1 DTD either by individual jurisdictions or by the CD 1.1 spec?
As far as the Court Document proposal goes, the aim is to offer a single
simplified CD 1.1 DTD (not multiple subsets) as something purely optional and
perhaps easier to get started with than taking on the full CD 1.1 DTD. The
object is to trim from the full CD 1.1 DTD elements and attributes that are
optional and not frequently used in most court documents, such as the
"notarization," "personIdentification" and "organizationIdentification"
elements, some of the optional attributes, etc. The assumption is that it
might be easier to start learning with a simpler subset of the full DTD. The
tradeoff is that a simplified DTD would be useful for validating basic, "plain
vanilla" XML court documents but not those requiring the richer markup in the
full DTD. Of course, maybe you are suggesting that simplifying the CD 1.1 DTD
won't make it any easier to learn and is not worth
pursuing.
Finally, are you concerned that some features of the
proposed simplified CD 1.1 DTD make it an improper subset of the full DTD? If
there are some specific "bloopers" in the simplified Court Document DTD,
please let us know what they are. I'm much more receptive to correction from
my friends than from strangers.
Thanks for sharing the comments.
Rolly
-----Original Message-----
From: Bergeron,
Donald L. (LNG)
Sent: Mon 9/9/2002 7:56 AM
To: 'John
Messing'; legalxml-courtfiling@lists.oasis-open.org
Cc: Bergeron,
Donald L. (LNG)
Subject: RE: [legalxml-courtfiling] Salt Lake
minutes
Thank you John for your posting. I think that it will go far
in increasing
understanding.
On the specific issue of subsetting
the DTD, I have a concern. The overall
LegalXML model as expressed by
Todd Vincent and reiterated many times is
that this dtd is over-inclusive
but optional to cover as many conditions as
practical and to not place
unnecessary constrains on any jurisdiction using
this dtd.
If
there is a way of always getting a proper subset of a dtd, and
validating
that condition well we may use it. Some of the approaches that
you mention
can provide part of this service. However, the part that it
is always a
proper subset is at risk as humans modify the parameter
entities on which
most of these are based. Only one method, the
Architectural Forms approach
has the concepts needed. This approach is
not present in most commercial
software nor is it likely to as it has
regretfully fallen out of favor. This
may be remedeiated W3C Schema
World, although I am not sure of this. This
constraint should be looked
upon as a requirement for CF 2.0. During the
implementation, if the
requirement can not be met we can address it as a
change control to the
requirements.
Beyond this, the constraints of the set has always been
a goal and
requirement of the court policy standards. This is only part
of the policy
requirements, but it is a part. One should think setting
constraints that
are jurisdiction specific in a way that is localized in
one place. This
allows the person developing the constraints to look at
the full constraint
set so that it can be balanced and tuned. When you
spread the validation to
two functions that are invisible to one another
you increase the very human
risk of getting it wrong.
This should
be further studied during the research needed to implement
the
requirements of court policy 1.0.
Regards,
Don
Bergeron - Chair - Court Policy
Sub-committee