Bookmap/XNAL repercussions

[Robert D Anderson <robander@us.ibm.com>]

As I make the bookmap changes from today's meeting, I'm also merging it
with the other 1.1 prototype files that are available. While doing that,
the decision to make XNAL into a domain has demonstrated some side effects.

The XNAL domain adds one element, authorinformation, as a specialization of
<data>. The <data> design has it appearing in many more locations than I
had first understood, including inside most other meta elements. This means
that, among others, the authorinformation element now shows up inside
elements like <source>, <category>, and <platform>.

I've mentioned before that the XNAL domain will not be useful in topic
specializations until domain specialization is enhanced, allowing you to
limit where your new element shows up. That enhancement has been talked
about for DITA 1.2. I thought that bookmap's constrained model worked
around this problem, but it is becoming clear that I was wrong. So,
primarily for usability reasons, I would propose the following changes to
bookmap:

1. Do not include authorinformation as a domain. Instead, include it
directly in the bookmap as a child of bookmeta.
2. Continue to define all of the XNAL elements in a separate module. In
DITA 1.2 or DITA 2.0, plan to move XNAL to a domain. This will be easy
because it is already maintained in its own module.

Are there any concerns with this proposal?

Robert D Anderson
IBM Authoring Tools Development
Chief Architect, DITA Open Toolkit
(507) 253-8787, T/L 553-8787