FW: [legalxml-courtfiling] Preliminary List of "Roles" in LegalProceedings

[John McClure <jmcclure@hypergrove.com>]

John,
My sentiments also, with an exception, my friend, to deferring this work to
V2.0.... For V1.0, I believe the present course -- to the extent it does not
adopt the W3C's Resource Description Framework from the outset -- is highly
inimical to achieving our organization's goals and objectives. The present
methodology we've adopted by default appears to me to be yielding a set of
haphazardly constructed DTDs for court records and court documents. No
information science is being applied at all, with the result that it is going to
be more costly to the judicial branch of governments to implement supporting
systems. That is annoying to taxpayers in general, and I believe directly and
undeniably diminishes our analytic abilities during our "war" on terror, and
beyond.

I don't mean to sound harsh, but it's difficult to continue to hear unjustified
conclusions against using the RDF as a guide in our work equal in stature to the
Extensible Markup Language itself. Both are formal W3C Technical
Recommendations, and both have been so for almost four (4) years now. The RDF
defines a sensible architecture equally applicable to commercial invoices as it
is to court records and court documents. RDF is a STANDARD from the principal
standards body we cite in our charter, so we should be reflexively desirous to
adopt it. LegalXML should also encourage the federal government to adopt RDF as
a standard, immediately, to truly improve data-sharing among its branches. RDF
is a STANDARD that should be adopted by all regulatory agencies, to improve the
quality of information solicited from industry, to minimize the cost and
maximize the effectiveness of the envisioned "data mining" across public and
private datasets that undergirds part of our present war strategy -- there is no
reasonable downside to standards adoption; there is no justification not to
adopt RDF, now. Simply standardizing on XML is hardly sufficient -- a babble of
grammatically dissimilar dialects clearly results.

There is also no published alternative architecture that I'm aware of that
compreensively addresses both instance documents and their metadata. XML Schema
only discusses metadata, and is almost silent on instance document structure.
RELAX-NG is a hack on XML Schema, and even seriously proposes a nonXML encoding
scheme. (Both of which are handicapped by being only incidentally
object-oriented). Understandably, the single specification that LegalXML has
built to date is "cautious" in its formulation, given its timing relative to our
understanding RDF, and given that WSDL is not yet a formal W3C standard.
However, nothing prevents our specifications from now building upon an
undeniably important standard established by the W3C -- RDF -- one that is a
documented part of a decent architectural vision, that is supported by
open-source and commercial tools, that is written about in the popular technical
press (books, magazines), one that is technically pretty darn good.

Perhaps if future specifications from LegalXML were guided by their own
architectural vision in some way, I'd be more encouraged. But none is published
or mentioned anywhere at any time. This is why I call the DTDs created to date
haphazard. This is why I think the conversation aimless and nonspecific. It is
one that has yet to draw any simple distinction between the requirements of
exchanging court records as opposed to court documents, as they do have entirely
different stakeholders interests to satisfy, and different technical approaches
to be taken.

I do know that you are one of the strongest members of this organization, that
you come to this table as a smart attorney with a hobby horse for XML markup.
Thankfully you're the (ideal) leader of the eNotary TC. Your expert
participation and good will are essential throughout this process of developing
work products used principally by technologists. So, I hope you agree that at
some point the process of designing the markup itself is best led by the
technologists who'll eventually have to eat this dog food. It shouldn't be a
goal here to "simplify" extant standards so that non-technologists could
presumably perform a job for which they are not trained, which they're unlikely
to do to any degree in any event. Likewise, I am not trained in law, but am here
to join forces with functional experts like yourself to achieve the goals that
we share, deferring to you in matters in which I have just a common man's
understanding. We are creating a technical product of central importance to
technologists, and I must stand opposed to seeing a quality product consistently
deferred for no given reason beyond a concern about political landmines. So, we
should adopt RDF now, not in V2.0.

Regards,
John

Hypergrove Engineering
Port Townsend, Washington

-----Original Message-----
From: John Messing [mailto:jmessing@law-on-line.com]
Sent: Thursday, November 21, 2002 6:03 PM
To: legalxml-courtfiling@lists.oasis-open.org
Subject: [legalxml-courtfiling] Preliminary List of "Roles" in Legal
Proceedings


My understanding comes more from contemplating personal and anectodal
programming experiences rather than formal training. My claim to expertise is
thus all but non-existent, but to me there a number of attributes like a role
which do not meet a necessary level or degree of persistence which I think is
implicit in a proper definition of a property.

Your question about multiple roles involves perhaps less case management
issues(relational database tables) and more courtroom formalities, strategies
and tactics, including legal arguments, precedents and authorities. One does not
normally describe what happens in a courtroom in  terms of roles (participants
are more interested in processes and results generally speaking). It is
difficult to give an honest answer about the existence and need for multiple
roles without a new discipline of thought emerging specific to this content
area. The RDF statement about the criminal law looks like it could adapt itself
more readily to such an undertaking than straight XML from what you have said.

I find the attempt at creating a hierarchy of entities to be noteworthy. Too
often it is tempting for the sake of expediency to include in a dictionary terms
in pork-barrel fashion so that a certain cartesian legal quality is immediately
compromised and like the saying about virtue itself goes, once compromised, it
is not easily regained.

Unfortunately, the particular hierarchy you have outlined from time to time
inadvertently steps through a minefield of semi-sacred terms and politically
correct behaviors with the grace of a bull in a china shop, but that is only out
of unfamiliarlity with the terrain, which perhaps is easily cured by obtaining
the services of a guide who knows the way.

I think you posting may be of considerable help to those looking to work on 2.0
and data dictionaries.

Best regards.



-----Original Message-----
From: John McClure [mailto:jmcclure@hypergrove.com]
Sent: Thursday, November 21, 2002 4:11 PM
To: legalxml-courtfiling@lists.oasis-open.org
Subject: RE: [legalxml-courtfiling] Preliminary List of "Roles" in Legal
Proceedings



Well, in the world of OOP, objects only have properties (other than a pointer to
their class metadata). Given this, a role is as much a property of the Person
object as much as title is. Lin's statement refers to the notion that in the
world of RDBs, one creates an intersection table between its Person table and
its Case table, each row of which contains the Role being played by the Person
in the case (as well as pointers to specific Case and Person rows). Of course in
XML we don't worry about tables per se but we still worry about the cardinality
of each property associated with an object. For example, Lin's system probably
doesnt allow multiple titles for a given individual, but I think one should, in
an XML encoding, because that is the reality -- people are able to claim
multiple honorific titles -- while most RDB systems store multiple honorifics,
comma-separated, in a single field (a huge no-no for XML encoding), because it's
not worth the effort to establish a separate table keyed by Person.

I think a key question is whether in fact multiple Roles can be played by a
single Person in a single Case. Since I am but a neophyte in case management
systems, I haven't a clue. If the answer is yes, but very infrequently, I still
take that to be a yes, and would recommend that markup allow such statements of
fact. Besides that, I'd suggest that the name of what's being discussed be
sharpened to include the event in which the role is being played by the actor.
For instance, CriminalTrialRole would identifiy all the roles found in the
generic criminal trial. By 'chunking' the discussion in this way, much benefit
and faster progress is readily assured. If one is using a metadata markup
language that provides multiple inheritance (like, RDF), then the same role can
be a subclass of multiple categories. If not (like, XML Schema), then there's a
considerable challenge ahead.

I substantially agree with Roger's explanations of the distinction between Role
and Title. The statement about a "Judge" -- that one owns the title throughout
it all -- needs much more refinement I believe. To me, a Judge is BOTH an
honorific AND an occupation (as defined by the US Government). Yes, any Judge
can be an ElectedOfficial or an AppointedOfficial but, as you might suspect, I
would establish these as sub-terms of a QualifiedOfficial term, applicable as
they would be to any subclass of Judge (as an occupation). Anyway, what is
crystal clear to me is that "Judge" is not an example of a Role. Rather, for the
role fulfilled by a Judge in a CriminalTrial, I would suggest that such be
called a CriminalTrialJudge, a term within the CriminalTrialRole category, which
is a subcategory of LegalRole (or CourtRole or JudicialSystemRole), which is a
subclass of Role, which is a subclass QualifiedActor. [Aside: a Judgeship is
another interesting item, because it refers to the jurisdiction had by a
judge...] For the honorific case, I wouldn't establish Judge as a subterm of
Honorific -- rather, "Judge" is an *instance of* an Honorific and I suggest
publishing such as a resource standardized by LegalXML, like <Honorific
rdf:ID='http://www.legalxml.org#Judge'><en>Judge</en></Honorific> which can be
referenced then by systems in a standard way such as <Person><addressedUsing
rdf:resource='http://www.legalxml.org#Judge'/></Person>. While I'm at this, for
a role, I would hardly recommend a <Role> tag, but rather a more readable, more
intuitive, more explicit, RDF treatment, as in <Judge><actedAs
rdf:resource='http://www.legalxml.org/dictionary.rdf#CriminalCourtJudge'/></Judg
e> for the simple, non attributed referential statement, and for the more
complex statements that are sure to occur:

<Person rdf:ID='Person1234567'>
	<LegalName><en>John Marshall</en></LegalName>
	<addressedUsing rdf:resource='http://www.legalxml.org#Judge'/>
	<employedAs rdf:resource='http://www.legalxml.dictionary.rdf#Judge'/>
	<actedAs>
	     <CriminalCourtJudge>
		<addressedAs><Honorific><en>Your Honor</en></Honorific></addressedAs>
		<at rdf:resource='http://www.USDistrictCourt/District5#Location12345'/>
		<occurred><ElapsedPeriod><iso>2002-12-12/P4D</iso></ElapsedPeriod></occurred>
		<forEvent rdf:resource='http://www.USDistrictCourt/District5#Case1234567'/>
	     </CriminalCourtJudge>
	</actedAs>
</Person>
<CriminalCase rdf:ID='Case1234567'>
	<occurred><ElapsedPeriod><iso>2002-12-12/P4D</iso></ElapsedPeriod></occurred>
	<at rdf:resource='http://www.USDistrictCourt/District5#Location12345'/>
	<presidedBy rdf:resource='http://www.USDistrictCourt/District5#Person1234567'/>
	<prosecutedBy
rdf:resource='http://www.USDistrictCourt/District5#Person7654321'/>
	<answeredBy rdf:resource='http://www.USDistrictCourt/District5#Citizen12345'/>
	<arguedBy rdf:resource='http://www.MaineBarAssociation#Attorney321'/>
	<decidedBy rdf:resource='http://www.USDistrictCourt/District5#Jury12345'/>
</CriminalCase>

Of course, this can be equally exchanged as

<CriminalCase rdf:ID='Case1234567'>
	<occurred><ElapsedPeriod><iso>2002-12-12/P4D</iso></ElapsedPeriod></occurred>
	<at>
	     <CourtHouse
rdf:about='http://www.USDistrictCourt/District5#Location12345'>
		<PhysicalAddress/>
	     </CourtHouse>
	</at>
	<presidedBy>
	     <CriminalCourtJudge
rdf:about='http://www.USDistrictCourt/District5#Person1234567'>
		<LegalName><en>John Marshall</en></LegalName>
		<addressedAs><Honorific><en>Your Honor</en></Honorific></addressedAs>
	     </CriminalCourtJudge>
	</presidedBy>
	<prosecutedBy>
	     <CriminalProsecutor
rdf:about='http://www.USDistrictCourt/District5#Person7654321'>
		<LegalName><en>Joe Prosecutor</en></LegalName>
	     </CriminalProsecutor>
	</prosecutedBy>
	<answeredBy>
	     <Defendant rdf:about='http://www.USDistrictCourt/District5#Citizen12345'>
		<LegalName><en>Jill Defendant</en></LegalName>
		<defendedBy>
		     <PublicDefender rdf:about='http://www.MaineBarAssociation#Attorney321'>
		     	<LegalName><en>Jane Attorney</en></LegalName>
		     </PublicDefender>
		</defendedBy>
	     </Defendant>
	</answeredBy>
	<decidedBy>
	     <Jury>
	          <consistedOf>
		<rdf:bag>
		<LeadJuror rdf:resource='http://www.USDistrictCourt/District5#Citizen0121'/>
		<AlternateLeadJuror
rdf:resource='http://www.USDistrictCourt/District5#Citizen0122'/>
		<Juror rdf:resource='http://www.USDistrictCourt/District5#Citizen0123'/>
		<AlternateJuror
rdf:resource='http://www.USDistrictCourt/District5#Citizen0124'/>
		</rdf:bag>
	          </consistedOf>
	     </Jury>
	</decidedBy>
</CriminalCase>

The whole POINT of a dictionary then is to identify the range of elements that
can validly be referenced by, or within, the "predicate elements" <occurred>,
<at>, <presidedBy>, <prosecutedBy>, and the others that "connect" two nouns.
Creating a hiearchical dictionary of terms is not merely an intellectual
exercise of passing interest -- it relates directly to the validation of XML
streams emanating from the judicial system. Now, it's been said that this seems
pretty wordy, but I daresay that such explictness should be an ongoing
requirement as more and more complex information structures are desired to be
transmitted in a reasonable, stable, fashion. Simply having a <Role> element
won't cut the mustard in the final analysis in my view.

For more information about this "style" of markup, please reference
1. "Make Your XML RDF-Friendly" at
http://www.xml.com/pub/a/2002/10/30/rdf-friendly.html
2. "RDF - What's it good for?" at
http://www.xml.com/pub/a/2002/11/13/deviant.html
3. "Ontology Building: A Survey of Editing Tools" at
http://www.xml.com/pub/a/2002/11/06/ontologies.html

Regards,
John

>-----Original Message-----
>From: John Messing [mailto:jmessing@law-on-line.com]
>Sent: Thursday, November 21, 2002 8:50 AM
>To: legalxml-courtfiling@lists.oasis-open.org
>Subject: RE: [legalxml-courtfiling] Preliminary List of "Roles" in Legal
>Proceedings
>
>
>Another way of looking at this may be that if there is a person
>object, a title is most clearly a property of the person object
>whereas the role is probably not.
>
>---------- Original Message ----------------------------------
>From: "Winters, Roger" <Roger.Winters@METROKC.GOV>
>Date:  Thu, 21 Nov 2002 07:40:27 -0800
>
>>As you have described it, then, does this follow?
>>
>>ROLE = something one enters and leaves. We all play many roles--father,
>>co-worker, civil servant, mechanic, mentor, penitent, child, caregiver, and
>>on and on. In a case, one may be the plaintiff but, on the stand, also a
>>witness; the next week, called for duty, a juror and, if a Judge, one owns
>>the title throughout it all.
>>
>>TITLE = something that sticks with one (unless revoked or modified). Being a
>>plaintiff is not to acquire a title, nor witness, nor juror. A medical
>>doctor who steps into (and out of) such roles would still be titled
>>"Doctor."
>>
>>Roger Winters
>>Electronic Court Records Manager
>>King County
>>Department of Judicial Administration
>>516 Third Avenue, E-609 MS: KCC-JA-0609
>>Seattle, Washington 98104
>>V: (206) 296-7838 F: (206) 296-0906
>>roger.winters@metrokc.gov
>>
>>
>>-----Original Message-----
>>From: Lin, Meng-chun [mailto:Meng-chun.Lin@usdoj.gov]
>>Sent: Thursday, November 21, 2002 4:03 AM
>>To: 'Winters, Roger'; 'Mohyeddin Abdulaziz'; 'Chambers, Rolly'
>>Cc: legalxml-courtfiling@lists.oasis-open.org
>>Subject: RE: [legalxml-courtfiling] Preliminary List of "Roles" in LegalPr
>>oceedings
>>
>>>From the programming point of view, the titles and roles are separate in the
>>federal courts' case management database.  A more common case is a retired
>>judge becomes an attorney and represents a client in a case.  The judge
>>retains his/her title, but plays different role in different cases.  Also,
>>under the party, there are different "party roles" in a case.
>>
>>
>>
>>
>
>----------------------------------------------------------------
>To subscribe or unsubscribe from this elist use the subscription
>manager: <http://lists.oasis-open.org/ob/adm.pl>
>