issue 20 (conditional processing) discussion results

[Michael Priestley <mpriestl@ca.ibm.com>]

Summarizing status of the discussion
on conditional processing from our telecon meeting on Monday  May
1st. Discussion is relative to the current proposal at http://www.oasis-open.org/committees/download.php/17329/IssueNumber20.html

The following interested parties were
on the call: Robert Anderson, Dana Spradley, Paul Prescod, Paul Grosso,
Erik Hennum, Bruce Esrig, Michael Priestley.

We agreed that:

- props and base will be specializable
attributes both
using the same tokenized syntax and generalization behaviors
(open: should rev be added to these, or be specialized off props? if added
as specialization, do we want to allow filtering based on rev? currently
only allows flagging) 

- specialization will be allowed to arbitrary levels, and use Robert's
format for generalization 

- ditaval actions against parent attributes will match against their children,
but not vice versa 
example: excluding platform="linux" will also exclude opsys="linux";
but exclude opsys="linux" will NOT exclude platform="linux".

- value hierarchies will not be modeled in 1.1, but we will attempt to
build a demo capability into the toolkit showing how it could work in 1.2
using maps and keys (open: if this approach not acceptable to external
stakeholders eg Deborah Pickett, we will take the value hierarchy proposal
separately to the TC for a vote). 

- element-specific attributes, and non-tokenized
attribute specialization, are deferred to a later release
 
- negative values will not be modeled in 1.1 

- open: how to reconcile color conflicts when an element is flagged by
two matches? 

Michael Priestley
IBM DITA Architect and Classification Schema PDT Lead
mpriestl@ca.ibm.com
http://dita.xml.org/blog/25