Meeting minutes 2013 03 26

Attendance: Robert Anderson, George Bina, Radu Corvu, Kris Eberlein, Jarno Elovirta, Jim Tivy

Item 1: DITA-OT 1.8 Milestone 2 release

Released March 25. Contains updates to doc build from Kris (new CSS for XHTML, minor topic updates, task labels). Also includes updates from 1.7.4 fix release.

Item 2: DITA-OT 1.7.4 patch release

Released March 25. Contains several fixes from Jarno.

Item 3: DITA-OT 1.8 Milestone 3 plans

• Website improvements: Kris has suggested several usability improvements to main DITA-OT page and docs. Some already included in M2 build (new CSS). Other changes include updated page on who contributes, easier access from main page to "who is contributing / how to contribute" info. Also, pages on contributing should link to minutes from these contributor meetings. George notes: participating in webinar, some people think they need to be OASIS members in order participate in DITA-OT, so should clarify that.
• Robert: new doc page on changes to content during build; begin work on PDF flagging. Checked in minor feature this morning for flagging, which continues use of exsl:node-set, but would like to remove that.
• Kris: Doing some doc cleanup this iteration, not sure how much will get completed.
• Jarno: Probably similar to last iteration - working on things in local fork. Right now, work on refactoring gen-list and debug/filter phases - code is currently a mess. Currently we go through files twice - this can't work if (for example) getting content from databases. This is done to support current Subject Scheme design - so this means changes around subject scheme. Should write up these changes in a wiki page. If we include in 1.8 - will probably need a new Ant target, something like "preprocess2", because we don't want to break old process.

Item 4: Other topics

• DITA 1.3: Just a note that design work is picking up at OASIS, but little solid to work on yet, unless we want to incorporate minor fix items early.

• Subject scheme: lots of discussion, lots of changes, see Jarno's plans above. Considering new feature to allow people to pass in a subject scheme to use (as an option, as a plug-in, whatever), rather than specifying it from within a map. Will be working on this over next few milestones. George wonders about this - today if a map references the scheme, you know everything. Much discussion - Kris and George both see it being far more common to specify the scheme in the map, and worry that these changes will stop that from working. Nothing written yet, but expect that any current capability will remain; may just be less efficient. Question about which would win - scheme specified as option, or scheme referenced from map. Jarno/Kris initial reaction is map. Robert feels the opposite; if I am using another author's map, I cannot change the map, but I can change the parameter when I build. George compares that to XSLT parameters (command line wins over value in the XSLT). Regardless - Jarno refactoring scheme code to improve it. Suggests a call exclusively for schemes - not only what are we changing/fixing, but what new features are we implementing? There are a lot of things that could change; among them:

1. Our implementation is not well written, inefficient and hard to understand. Would be good to clean up the code, even if nothing else changes.
2. Our implementation likely details some items that are part of the specification (holes in our code).
3. The specification is incomplete for many items. For example, OASIS discussion item for today, if 2 ennumerations are specified for one attribute, which wins? For these, OASIS should clarify intent before we try to solve any problem.
4. New features can be added that may be very useful to some, but unrelated to the spec. For example, ability to specify scheme with command line or plug-in: demonstrated use case for some teams, irrelevant for others, but either way unrelated to the specification.

Robert will look for a time to have a call exclusively about subject schemes, probably in a couple of weeks.