Further conference calls will be scheduled for the Mondays of 1997-02-10 and 1997-02-17 at 14:00-15:00 UTC as usual. Note that the latter is Presidents' Day in the U.S. EM said that this might cause her problems, but did not object to the schedule.
On-line copies of subcommittee documents may be found at http://www.uic.edu/orgs/tei/trc/core/. (Enter this URL directly in your browser: nothing links to it, so you cannot navigate to it from another page at the same site.) Plain-text, HTML and TEI P3 versions of most documents exist as .doc, .htm, and .tei files respectively.
The group agreed to review TEI TRC Core W02, Summary of status: Problems assigned to the TEI TRC Core Subcommittee, which had been distributed by CMSMcQ. Items listed as having no disposition were discussed first, followed by items flagged as requiring further discussion.
[Here, and in section 2, I have reinserted the item numbers which appeared in W01, but which were omitted in W02 -- DD]
The group was of the opinion that this behaviour was due to user error.
ACTION: None
STATUS: Closed
Disposition documented in EM's addendum to notes for meeting of 1996-12-23:
Part 1: ACTION: EDITORS: allow figure on title page
Part 2: ACTION: LB will contact Fantz Wiering for clarification.
Part 3: No action can be taken. In the case of isolated notes within text, it may be possible to use entities to indicate the different notes. However, TEI cannot possibly do better than SMDL, and is not in a position to devise new markup schemes for things like music in any thorough way. The <FIGURE> element has a notation valued attribute that would allow SMDL to be used in an SGML document.
ACTION: EDITORS to amend default value of type attribute for <stage> to "mixed".
STATUS: Closed
The group discussed as many of these as possible in the time available, leaving the remainder to be discussed at subsequent meetings.
The group decided that the class of "asynchronous things" should be global inclusions. In addition to <addSpan> and <delSpan>, this class should include <span>. (<anchor> is already part of the class.)
ACTION: Add <addSpan>, <delSpan> and <anchor> to class %globincl.
STATUS: Closed
On reflection, the group decided that there were no deeper issues to discuss.
ACTION: Use spelling teiCorpus.2
STATUS: Closed
Again, the group decided that there were no deeper issues to discuss.
ACTION: Use spellings participant and participantGrp
STATUS: Closed
MSMcQ suggests in document W02 that %inter should contain no base-specific elements. The group agreed that this was a cleaner solution than noting that anything added to %inter should also be added to %common. LB pointed out that %common was hardly documented anyway, and would perhaps better be totally hidden. MSMcQ further proposed that %inter should function in the same way as %chunk in the core. (This would involve splitting %inter.) The proposal was accepted.
ACTION: Remove all base-specific elements from %inter; %inter should function in the same way as %chunk in the core; note in reprint of P3 that %common should not be extended; hide %common from view in a subsequent edition.
STATUS: Closed
The group agreed, LB dissenting, that empty lists should be allowed. Additionally, empty paragraphs -- useful as place-holders -- should be allowed. These changes would aid in the use of the TEI DTD of document production -- an application not originally envisaged, but which had turned out to be important. DD's earlier suggestion that a "tight/loose switch" on the DTD "would be nice" should not be acted on for the moment.
ACTION: Allow empty lists and paragraphs
STATUS: Closed
In response to the original complaint, the group was of the opinion that <s> should be allowed as a parent of <q> (and also of <quote>). More contentious were the wider issues of how this should be achieved, and of how many similar but currently-disallowed nestings should be permitted. (For example, why can <note> not appear in some elements while being allowed in some children of those elements?)
MSMcQ had circulated document WO3, "Note on element classes". There was much discussion of the relative merits of over-generation (allowing most nestings, even those seeming at first sight to be meaningless, useless, or unlikely ever to be needed) and helpful prescription (allowing, on a case-by-case basis, nestings which are currently prohibited, but which have been proven by experience to be desirable). No decision was reached, but the following possible courses of action were proposed:
In connection with the last option, the decision matrix is (or should be) a function of three binary variables: can the element occur between paragraphs?; can the element occur within paragraphs?; and is the element able to contain elements in class %inter? It is the last of these that is problematic.
Here's the matrix, and a cursory shot at populating it:
| Can occur between paragraphs? | ||
|---|---|---|
| Can occur within paragraphs? | No | Yes |
| No | (invalid) | %chunk |
| Yes | %%paraContent | ? |
| Can occur between paragraphs? | ||
|---|---|---|
| Can occur within paragraphs? | No | Yes |
| No | (invalid) | ? |
| Yes | %specialPara | ? |
(This corresponds to the section "Other Problems" in document W02.)
The following issues arose from discussion of other items, and should be considered at a subsequent meeting.
<teiCorpus.2> [that is the spelling we decided upon, isn't it?] cannot currently nest, but there is a feeling that it should be allowed to, in order to provide for corpora of corpora.
LB feels that the Recommendations should provide a standard method for handling this common requirement, either with specific elements, or by saying something like "use a <list> having a single <p> in each <item>".
Dominic Dunlop