=================================================
Minutes of the OASIS Lightweight DITA SC
Monday, 23 January 2017
Recorded by Keith Schengili-Roberts
In attendance: Rahel Bailey, Don Day, Kristen Eberlein, Carlos Evia, Mark Giffin, Tim Grantham, John Hunt, Jonathan Piasecki, Michael Priestley, Keith Schengili-Roberts,
Birgit Stackenbrook
Regrets: Scott Hudson
Previous minutes provisionally approved "pending quibbles" from Kris [Note from Keith: I have still to hear any further details about these quibbles].
==Status of the Committee Note==
Draft #3 of Committee Note (CN) has been circulated. The new version includes a revamped table containing elements and attribute equivalents across XDITA, HDITA and
MDITA.
Carlos' opinion is that the CN—minus the section relating to template specialization—is now close to being ready for circulation with the Technical Committee. The
template specialization section needs significant work, and there are holes elsewhere, but “real progress has been made”.
==Concerns About Template Specialization==
Kris circulated an email on January 19th that says while everyone wants specialization to be easier than it is now, she wasn’t yet convinced that the new attributes
and elements being proposed for LwDITA were needed. She also expressed concern that Michael has a vision for template specialization that has yet to be fully fleshed out in a documented form.
If we want to consider a release of LwDITA that does or does not include template specialization, then it either needs to be in the CN, or it should be left for a
subsequent release.
Michael briefly explained how a template-based specialization could become portable so that any topic becomes portable across systems, and can be used by other toolchains.
He also stated that most of the ideas for template-based specialization are covered in the slidedeck from the DITA NA 2016 conference.
Tim responded that he encountered some open issues that the presentation from Michael did not address, but otherwise he has not had any problem understanding what
was there. Mark concurred.
Kris further emphasized that the SC will eventually need to define the LwDITA specification in a manner that effectively provides its algorithms to toolmakers. As
Tim commented: “the specification must be toolable".
To Do:
Michael to meet with Mark and Tim on Friday, January 27th 2017 to flesh out the ideas around the template
specialization ideas for LwDITA. A draft section on this will be made available for the next meeting
Resolved:
Based on this draft, the SC will decide to either incorporate template specialization with the CN or not.
To Do: Michael also offered
to talk about template specialization to the TC on February 7th. Kris will take this to the TC.
==Issues Relating to LwDITA Open Questions==
Tim asked whether LwDITA should attempt to validate the content. As an example of what he meant by this, he mentioned that Birgit had an example where she opted for
creating a choice model from sequences, so LwDITA cannot simply ban the creation of choice models when derived from a sequence; the problem is that this sort of thing is hard to incorporate within a validation mechanism. He concluded by saying that in “full”
DITA it is up to the user to ensure that the specialization is valid.
Michael responded by saying that we could “hard-code” of these ideas in the eventual specification. The focus would be on what error conditions we need to prevent.
The only edge cases identified so far are in-topic, since we are only working with a single level of specialization. Michael also suggested that another way to approach this problem is to punt it in the same manner that “full” DITA does, and leave it up to
the specializer to not break the system (even though they of course they can). In response to this last point Birgit asked whether we could get more guidelines to ensure that people do not break the specialization.
Another issue Birgit and Tim encountered: with the dl element there is currently no way to set the number of its occurrences to more than one. Michael responded that
we may need an equivalent rule we need to handle this.
A third issue: how should shortdesc be handled in MDITA and HDITA? Should it just be the first paragraph by default? This is still an outstanding issue.
To Do: Tim to send a refresh
of any outstanding issues he has encountered to the email list.
Continuing Work:
Work
on the next draft of the introductory LwDITA committee note.