[This local archive copy is from the official and canonical URL, http://www.xmlxperts.com/minutes1118.htm; please refer to the canonical source document if possible.]

ISO 12083 Minutes Nov. 15 & 18, 1998, Chicago

Minutes:
ISO 12083 Meeting (WG6)

Chicago, IL.
November 15 & 18, 1998

Roll and introductions began the program on the 15th. Attendees included:

Review Minutes from last meeting

Dianne Kennedy read the minutes of the last meeting.

Review of Status of the Standard

Pat Harris briefly reviewed the status of ISO 12083. The standard, dated 1994 is up for formal review next year. It is appropriate that this working group update the status. Ms. Harris will review the ISO rules and advise the working group as to the best procedure to follow. In the interium, the XML version of the standard seems to be the most pressing issue. It is this issue that must first be addressed.

XML DTD

This work began with a review of the differientiators between XML and SGML by Dianne Kennedy. Evan Owens pointed out that decisions must be made that will most likely make the XML version incompatible with existing SGML data. The group reached concensus that this was unavoidable. Since it seems clear that technology trends will mandate XML compatible data, providing guidelines and a target seems not only appropriate but responsible.

The remainder of the session was spent converting the article DTD into an XML article DTD. Along the way the following decisions were taken:

Use of the AND Connector

The current ISO 12083 DTD makes use of the AND (&) connector. This is not allowed in XML. The group revisited the requirement for the & and found that that the use of & was inconsistent. That is, appears that & models were provided when committee members felt strongly that they presented article data in a different sequence and could not agree on one model. Analysis showed that there were many other instances where an & model would be appropriate as well. Secondly, at the time the & was used, there was no standard method of transformation to enable users of 12083 to affect the output order of elements (in print, or on the screen for example). With DSSSL and XSL, this is no longer true. Most publishers use transformations routinely in their day to day work. Hence it was decided to eliminate the & models and "freeze" content models in a standard way. The following models were changed:

First Name / Surname Model

<!ENTITY % m.name
"((fname? & surname), (degree|school)*,
role*,
(street|city|state|country|postcode|san|email|postbox|phone|fax)*,aff?)"
>

changed to

<!ENTITY % m.name "((fname?, surname), (degree|school)*, role*, 
(street|city|state|country|postcode|san|email|postbox|phone|fax)*,aff?)">

Figure Group

<!ELEMENT figgrp - - (title? & fig*) >

changed to

<!ELEMENT figgrp - - (title? , fig*) >

Exclusions

Exclusions are not allowed in XML. Hence a decision had to be made about the exclusions. Upon study it was determined that the exclusions in ISO 12083 were to exclude those elements defined by the parameter entity %i.float from content models that contained % i.float, thus precluding floats within floats. There were 2 approaches to remedy this problem. First exclusions could simply be eliminated from the DTD. It would require the "user beware." That is, the user would have to take responsiblity for the proper use of elements. The second approach would be to prevent the use of floats within floats by creating quite a number of new elements that did not include %i.float in certain situations.

The working group agreed that the "user beware" approach was the most reasonable. It would complicate the DTD to add a host of new elements just to prevent the use of floats within floats. So exclusions were simply eliminated from the DTD. This would not make existing data invalid. Below is one example of this change:

<!ELEMENT footnote - - (no?, %m.pseq;) -(%i.float;) >

changed to

<!ELEMENT footnote - - (no?, %m.pseq;) >

Inclusions

Inclusions were used in the ISO 12083 DTD at a very high level to include the floats (%i.float). XML does not allow inclusions. So the group worked through the DTD to determine where floats really should be allowed. The group determined that the logical placement for "figgrp, footnote, note" were wherever #PCDATA was allowed. Note that this may make some existing SGML data invalid. This is true is floats were placed between elements. However with new transformation standards and tools, this was determined to be the best course to follow. Hence a new parameter entity was developed and placed in the DTD wherever #PCDATA remained the sole content model. 

<!ENTITY % p.pcdata.zz "#PCDATA|%i.float;|%p.em.ph;" -- added for pcdata -->

So . . . 

<!ENTITY % m.date #PCDATA      -- date -->

changed to

<!ENTITY % m.date   "%p.pcdata.zz;" -- date -->

Mixed Content

XML has a very prescribed mix content model which requires that #PCDATA be placed in an OR group with other elements following. ISO 12083 did not conform to that rule. The changed required examination of each parameter entity where #PCDATA was used and elimination of some sub groupings created by the use of parens. This change should not affect the validity of existing SGML data. Groupings were also reviewed to placed first in the OR group as per XML rules.

November 18 Meeting

Attendees on November 18th included:

A special interest meeting was held the evening of November 18th. At this meeting the results of the working session earlier in the week was reviewed. Attendees agreed with the proposed approach to creating an XML version of ISO 12083. Two items were stressed by the group. First that the work on the DTDs be taken forward outside the meeting sessions and be made available for review. Second that the work be posted on a Web site for comment.

Summary

Dianne Kennedy will prepare XML versions of each DTD in ISO 12083. Bugs and updates will be gathered and incorporated into the DTDs along with comments gathered on the XML versions of the DTDs during the ISO 12083 meeting in Granada Spain in April 1999.

(Return to TOC)