[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [dita] 1.2 Requirements Ranking
Hi, Eliot:
Good to have this thoughtful review of the full list. A specific follow up on issue 12008:
- 12008 - Potentially very complicated, not clear that it's needed, could be controversial
The proposal (http://www.oasis-open.org/committees/download.php/14936/Issue34.html) cites some posts on the user group and TC mailing lists that provide background but doesn't itself provide an adequate explanation of the two most common scenarios. So, here goes:
Scenario 1: Some adopters want to simplify content models without adding new semantics.
For example, to exclude phrase elements and text (allowing only block elements) from <section> and specializations with a content model where that constraint could also apply (such as <context> and <prereq>). In particular, Paul Prescod surfaced that issue to the TC early on.
Scenario 2: Some adopters want to provide specialized domain elements as an alternative or replacement to the base element in specific content models.
For example, the style policies stuff that John Hunt and I are cultivating needs to add <blockLayout> as an alternative to <data> within <p> and its specializations, <linkLayout> as an alternative to <data> within <xref> and its specializations, and so on. If <blockLayout> and <linkLayout> are available in all <data> contexts (so that, for instance, <linkLayout> is available within <p>), the document type becomes unusable.
If the full constraints proposal seems too much for DITA 1.2, one possibility would be to support contextual domains only as a partial step down that path so at least the second scenario can be supported.
One additional case for trying to make progress on contextual domains: the control of the document type shell over the nesting of topics resembles contextual domains, so it might be plausible to recognize both with a single mechanism introduced as part of topic-domain unification. Such a mechanism could, for instance, allow runtime detection of incompatibilities between shell documents that use the same topics but impose different topic nesting.
Hoping that clarifies,
Erik Hennum
ehennum@us.ibm.com
"W. Eliot Kimber" <ekimber@innodata-isogen.com>
03/20/2007 08:51 AM |
|
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]