Re: [legalxml-courtfiling] Recap of person object discussion; stillneedinput

[John Messing <jmessing@law-on-line.com>]

So there is no further mistake, let me repeat for a third time that I favor the GTRI proposed person object as a superior technology to the linear relations of the present actor element. 

I also oppose the proposed response from Greacen Associates because it ignores a carefully stated invitation by GTRI to explain how a hierarchical inheritance model of a legal entity and person object could work. I think it is a reasonable request and a realizable goal, given some study and effort.

I therefore oppose any rush to judgment, either by outright rejection or unthinking acceptance, of the GTRI proposal. I think it unwise to take a firm position on the GTRI person object as it has been presented without careful study and reasoned discussion within the TC under Oasis rules, regardless of any reasons that may be advanced to the contrary.

Because I do not now choose to controvert any of the other statements or assertions of the prior posting, including but not limited to those about alleged agreements reached in Salt Lake City, one should not assume I am in agreement with any of them.


Very truly yours.

Best regards.

John Messing

---------- Original Message ----------------------------------
From: "John M. Greacen" <john@greacen.net>
Date: Mon, 02 Sep 2002 03:46:01 -0600

>Let me attempt to moderate this discussion, answer some of the questions
>that have been raised, and, once again, ask for input on the critical
>issue(s) facing us.
>
>First -- My posting is not a proposal to modify Electronic Court Filing
>1.1!!!  It has nothing to do with 1.1.  It has to do with the Global
>Data Dictionary Schema effort, which we agreed in Salt Lake to
>incorporate in our version 2.0.
>
>Second -- Because we will be heavily affected by the Data Dictionary
>when we go to 2.0, we have to give our input now.  The object oriented
>Data Dictionary Schema is being developed now -- on a very tight time
>schedule.  You designated Robin Gibson and me to be your delegates to
>the Data Dictionary development effort.  We are asking for your help in
>performing our role.  We do not have the luxury of sitting back, making
>no comment, and saying we will or will not go along with whatever the
>Data Dictionary group comes up with when we get to 2.0.  The Data
>Dictionary Schema is being designed so that groups like ours can
>incorporate large pieces of it by reference in our future schemas.  It
>will be a highly influential structure, which we will not be at liberty
>to ignore.
>
>Third -- How did it come to pass that our "actor" and "role" elements
>did not get included in the current data dictionary?  The reconciliation
>effort addressed elements used in two or more of the existing XML
>standards, plus elements in one standard that other groups were
>interested in.  "Actor" and "role" appear only in Court Filing;  they
>are not in the Rap Sheet, RISS, or AMVAA.  None of those groups
>expressed any interest in the "actor" or "role" concepts.  They had no
>objection to our continuing to use them but they thought they had no
>need in their businesses for a concept broader than "person."  There are
>many such elements in the standards of the four groups that were not
>included in the data dictionary.  In short, the current data dictionary
>is not a compendium of all elements contained in the four standards that
>were reconciled; it is a subset of those elements -- those used in more
>than one standard plus those of interest to more than one standard
>setting group.
>
>Fourth -- From the small amount of discussion that has focused on the
>actual issue, I gather the following:  Rolly prefers the person object
>model over our "actor"/"role" model.  Gary preferes "actor"/"role".
>Both have worked extensively with the actor/role structure.  Rolly did
>not like it for Court Document;  Gary does like it for the work that he
>is doing outside court filing.  No one else has made an argument one way
>or the other -- or I have missed it.
>
>In a telephone conversation with Rolly, we came up with this additional
>consideration:
>--  The use of actor/role requires that each court specify in Court
>Filing Policy the roles that it requires and supports - e.g., (for
>criminal cases) district judge, magistrate judge, justice of the peace,
>assistant district attorney, assistant public defender, defense counsel,
>witness, defendant, juror, confidential informant, victim, etc.  For
>every additional case category, additional, different roles are needed,
>e.g., (for domestic relations and domestic violence cases) petitioner,
>respondent, child, grandparent, relative, intervenor, child protective
>services attorney, assistant district attorney, guardian ad litem,
>domestic violence victim, counsel for petitioner, counsel for
>respondent, counsel for intervenor, etc.
>--  Conversely, a comprehensive person object seems to contemplate the
>delineation of every one of these different "person types."  Is that
>possible?  Is it workable?  Different jurisdictions call the same person
>type by different names, e.g., district attorney, prosecutor, county
>attorney, prosecuting attorney, special prosecutor, assistant county
>attorney, deputy county attorney, supervising deputy county attorney.
>How does this fit within the "person object?"  For the object to be
>complete, must it include most if not all of these "person types."  GTRI
>expects that the model will be extended to include types not specified.
>How do we handle types specified but called something else in a
>different jurisdiction?
>
>I have received a response to my inquiry whether the "super object" in
>the proposal is intended to be our "actor" element.  It seems not.  I
>did not hear opposition to an "actor" or "entity" object at the Denver
>meeting -- but I had to leave early and it may have been expressed after
>I left.  Here is the question I asked, and the answer I received, set
>forth below in italics.
>
>Webb -- In the Electronic Court Filing 1.0 and 1.1 specifications, we
>chose to
>use an "actor" element, which includes, but is not limited to "persons."
>The
>courts and lawyers need to be able to include entities such as
>corporations,
>partnerships, and associations. We also need to include "things" in the
>same
>way in some instances, such as cases in which an automobile or a large
>amount
>of currency is proposed for forfeiture and court rules require that the
>automobile or the money be the "party" in the case. Do you intend that
>the
>proposed "SuperObject" will serve this purpose? If so, is it necessary
>at
>this point to define the "SuperObject" and its other components?
>
>I have referred the proposal to other members of our Technical Committee
>and
>may have other comments to submit when I hear back from them.
>
>John,
>
>In earlier versions of our design, we included an "actor" class, from
>which we derived organizations and persons.  However, that seemed to be
>controversial at the meeting in Denver, so we backed off and stuck to
>modeling a person.
>
>XML schema is limited to a single inheritance model, and it is a
>difficult shortcoming, which restricts us from using the inheritance
>model to support this type of polymorphic substitution.  RDF supports
>multiple inheritance, but until that becomes widely supported and
>acceptable, we will likely have to live with the limitations of XML
>schema.  Developing an alternative standard method for supporting the
>level of substitution you desire would be a
>Good Thing, and any input is welcome.
>
>With the limitations of XML schema, you could get a similar effect to
>that substitution using the "choice" construct of XML schema.  Creating
>an "actor" type to perform that substitution seems contrived, and is
>probably not advisable.  With a single-inheritance model, we should
>limit subtyping to cases where there are common data components among
>the supertype and the
>subtypes.
>
>Thanks,
>The GTRI team
>
>Folks, Robin and I need guidance on this issue.  Comments on this
>proposal closes on September 2d.  Based on what little I have heard, and
>my own thinking, I am leaning towards the following sort of response:
>
>"I have circulated this for response within the Electronic Court Filing
>Technical Committee.  I received one favorable and one unfavorable
>response.  However, based on my understanding of the needs of courts and
>lawyers, I oppose the current person object as being too narrow to serve
>our business needs.  Our scope goes well beyond criminal cases.
>However, both in criminal and civil cases, we need to be able to handle
>persons, organizations, and things as parties to and participants in
>lawsuits.  An "actor" or "entity" object would meet our needs.  I also
>believe that such an object would fully meet the needs of law
>enforcement organizations that care only about "persons."  An "actor" or
>"entity" object could have three subobjects of "persons,"
>"organizations," and "things" or "objects."  The current proposed object
>does not support our business needs.  I also do not believe that it
>meets the needs of law enforcement, which must investigate crimes
>committed by corporations and other types of organizations, must
>interact with other organizations, and which has to deal as evidence
>with the "things" that end up being the subject of forfeiture cases.  I
>have not been provided with any suggestion for how our needs will be met
>by this or any other object.  Consequently, I do not believe that this
>object is ready for adoption until our broader category of business
>needs is addressed.
>
>One of our members has pointed out that the proposed object is
>inconsistent with the current data dictionary in several regards, for no
>apparent reason -- PersonalIDNumber has become PersonalID and
>AssignedID.  PersonType has been split into BioMetricIdentity and
>PersonalIdentity.  PersonDescription has become PersonalIdentity.  And
>the proposed object has no reference to name type (e.g., current, alias,
>doing business as, formerly known as, etc.).  I was under the impression
>that our charge was to make as few changes to the current data
>dictionary as possible.
>
>Finally, I am not clear on how detailed the person object needs to be in
>defining all of the person types that are used in the justice arena.
>For instance, I have come up with the following list for criminal
>cases:  district judge, magistrate judge, justice of the peace,
>assistant district attorney, assistant public defender, defense counsel,
>witness, defendant, juror, confidential informant, victim, etc.  For
>every additional case category, additional, different person types are
>needed, e.g., (for domestic relations and domestic violence cases)
>petitioner, respondent, child, grandparent, relative, intervenor, child
>protective services attorney, assistant district attorney, guardian ad
>litem, domestic violence victim, counsel for petitioner, counsel for
>respondent, counsel for intervenor, etc.
>
>Further, different jurisdictions call the same person type by different
>names, e.g., district attorney, prosecutor, county attorney, prosecuting
>attorney, special prosecutor, assistant county attorney, deputy county
>attorney, supervising deputy county attorney.  How does this fit within
>the "person object?"  For the object to be complete, must it include all
>of these "person types."  How do we handle types specified but called
>something else in another jurisdiction?"
>
>I continue to seek input.  Perhaps this specific possible response will
>help to focus your attention on the real issue(s) facing us.  Please
>respond no later than September 3.  Since the 2d is a holiday, I intend
>to invoke standard court rules allowing a response on the next business
>day!
>
>