Received: from UICVM (U35395) by UICVM (Mailer R2.07) with BSMTP id 0846; Thu,
 04 Jun 92 10:44:37 CDT
Date:         Thu, 04 Jun 92 10:15:47 CDT
From:         "C. M. Sperberg-McQueen" <U35395@UICVM>
Organization: ACH/ACL/ALLC Text Encoding Initiative
Subject:      further note and correction on ODD DTD
To:           "Robert A. Amsler" <amsler@starbase.mitre.org>,
              Susan Hockey <hockey@zodiac.rutgers.edu>,
              Nancy Ide <IDE@VASSAR>,
              "Donald E. Walker" <walker@bellcore.com>,
              Antonio Zampolli <GLOTTOLO@ICNUCEVM>,
              Steve DeRose <sjd%ebt-inc@uunet.uu.net>,
              David Durand <DGD@cs.bu.edu>,
              Harry Gaylord <GALIARD@let.rug.nl>,
              Rich Giordano <rich@r1.cs.man.ac.uk>,
              Dan Greenstein <DIGGER@UDCF.GLASGOW.AC.UK>,
              "D. Terence Langendoen" <LANGENDT@ARIZVM1.CCIT.ARIZONA.EDU>,
              Elli Mylonas <ELLI@HARVUNXW>,
              Allen Renear <ALLEN@BROWNVM>,
              David Robey <robey@vax.oxford.ac.uk>,
              Peter Robinson <PETERR@VAX.OXFORD.AC.UK>,
              Gary Simons <gary@TXSIL.SIL.ORG>,
              Syun Tutiya <Tutiya@russell.stanford.edu>,
              "David T. Barnard" <barnard@qucis.queensu.ca>,
              Michael Sperberg-McQueen <U35395@UICVM>,
              Lou Burnard <LOU@VAX.OX.AC.UK>,
              David Stanfield <u49112@uicvm>,
              Wendy Plotkin <U49127@UICVM>
 
You should all have received the promised sample sections of chapter
23, with Lou's voluminous commentary.  Since you all also have output
from the sections in question, you can compare the input with the
ultimate output.  If you would like copies of the originals without
the explanatory comments, please ask Wendy Plotkin to send you files
P2231 ODD, P2232 ODD, and P223REF EXAMPLES.
 
Questions are welcome; please post to Lou and me.
 
Several slips eluded our checking of the examples, however.  Some are
simple SGML errors occasioned by shortening the examples from the
text distributed to you, which had been validated.  A couple are slips
of the pen.  None are of great importance, so if you are not interested
in all the details of the DTD you may stop reading now.
 
Those I've noticed so far are:
 
Slips of the pen:
   - an ordered list has TYPE=ORDERED not TYPE=NUMBERED
   - the USAGE attribute on the TAGDOC element has five values not four:
          * req (required unconditionally)
          * mwa (mandatory when applicable:  if the feature occurs, or
            the information is available, tag and content must be
            supplied)
          * rec (recommended)
          * rwa (recommended when applicable; there are situations in
            which the use of the tag may not make sense, or may be
            optional; often this value has the sense 'it is recommended
            that you do it this way if you are going to do this kind
            of thing at all')
          * opt (strictly optional)
   - the ID attribute on the ATTDEF is strictly optional, not required;
     it is not used by any of the ODD processing software at present.
 
SGML slips:
   - the comment containing the string 'the ID value' needs a double
     hyphen at its end, not a single hyphen
   - strictly speaking the dtdFrag containing the P tag should have
 
          <dtdFrag id=d2231 n='Paragraph'>
 
     since 'Paragraph' is not a valid name in the syntax we are using
     (name too long).
   - exemplum can contain EG elements and P elements; comments on an
     example must be inside a P, rather than appearing directly in the
     EXEMPLUM as shown.
 
As these items show, Lou and I manage to get by quite well without
having all the details of the DTD in our heads.  (When we have a
question, we check the DTD.)  As long as your intention is clear, details
of SGML validity are unlikely to cause any problems in editing.
 
Michael