Meeting minutes 2016 11 03

Attendance: Robert, George, Radu, Jarno, Kris, Bill Gamboa, Eric Sirois, Meegan Tomlins, Roger Sheen, Eliot Kimber

Item 1: Any updates about prior releases?

Any concerns about 1.x, 2.x?

Item 2: DITA-OT 2.4 status and updates

Next major release - creating branch end of this week, no more features / fixes expected after that. Next week for any doc updates / release notes.

Jarno noticed that some of integration tests are failing, not noticed previously because driver class had some bugs that were skipping the errors. Fixing those now. Would be nice to have the tests working, after all, even if we don't have that many.

Biggest concern from Roger at this point - FOP 2.1 performance problems.

• Roger has seen significant slow downs with 2.1 which may be a concern for our users - our PDF used to complete in under a minute, now taking 5 minutes after the font cache is rebuilt, and but the initial build is 15 minutes. This will have a major impact, particularly for anybody with a continuous build env that destroys the old environment with each rebuild - they will see the font cache problem every time. Roger does not see any changes to the PDF that would make this worth the slow down, so may want to consider going back to older version.
• Jarno suggests comparing 2.1 (currently in build) to 2.0; if both are slow may want to down grade to version in DITA-OT 2.3. Has not seen such a slow down on his machine.
• Roger points out - it collects all fonts on the machine, a design heavy machine with 100s of fonts will see much worse impact. His own builds, same machine, are taking 15 min (with rebuild) compared to 1 previously; 5 compared to 1 after the font rebuild.
• Jarno did some quick comparison testing with 2.0 during this call, which seems to take just as long with rebuilding font cache. Overall - one goal with the move to 2.0 was that we would no longer store FOP in our library, but the dependency chain for 1.1 is broken, so we'd need to put the whole thing back in. Don't want to do this, but don't see a better option.

What's the big "selling point" of 2.4 in general - highlight items in release notes, features to talk about at DITA-OT day? Other than the general issue of collected fixes? Roger points out - new long form command syntax, passing parameters without the "-D" syntax. Also new coderef option that pull in content based on file content.

Jarno thinking of this as the release that stabilizes 2.X so we can move forward to 3.0.

Item 3: Doc updates and plans, update on documentation call

Roger has updated banner on dev docs, as discussed at last call. Allowed new feature additions without worry of confusion.

Changes to contribution page (new DCO process) also updated, so you can see those - much more up to date and cleaner. Left the project management topic, still needs some cleanup, but other stuff in there may be of use.

Item 4: DITA-OT Day 2016

November 13 in Munich. Are you ready?

"Reserved" session now to be filled with a talk about testing - describe the process, ideally solicit some feedback / suggestions and maybe some help.

New supporter: Antenna House.

Looking at ways to broadcast sessions into webinar that is less intrusive than in the past. Probably have a computer with webcam pointed at slides on screen, so that we do not need to have every presenter join the webinar and present. Tested yesterday and didn't see any issues. Webinar info now updated on site, need to promote it.

Item 5: Other topics

Discussion of testing leads to discussion of vendors testing DITA-OT in the context of their product, when upgrading. Some just upgrade, do not test, any problems come out later. Suggestion made that some vendors do not upgrade (or do so rarely) - seen when called in to consult on processing enhancements, but vendor product requires that plugins / enhancements work with much older releases.

George has interesting suggestion, and accidentally signs up for fascinating lightning talk. At CMS / DITA North America he showed off a "resolved DITA" transform, suggested for use with translation (somewhat controversial reception). This same plugin could be used to take advantage of the latest DITA preprocess, then chained to an older release that is used purely for formatting. Use DITA 1.3 content => process to "resolved DITA" with latest toolkit, getting all expected key resolution, etc => process with 1.X that supports special output extensions (could even skip preprocess after gen-list and debug-filter) => get output, using all DITA 1.3 features, but your old plugins keep working to generate the output you need. We look forward to this lightning talk at DITA-OT Day 2016.

Item 6: Backlog discussion

N/A, out of time...