See minutes online for a more detailed record of the discussions.
Tzviya gave a short overview of DPUB ARIA 1.0: the ARIA group (together with the DPUB IG) took the EPUB SSV terms, filtered them, and put the most useful terms in the ARIA vodabulary. DPUB ARIA 1.0 is a Rec, complete with a mapping of the terms to the accessibility engines. The question is now: do we need a 2.0? Is what exists good enough?
The subsequent discussion made it more clear that there is a difference between semantic types (body matter, back matter, indexes) and the accessibility features. ARIA is primarily for accessibility, ie, this distinction is important. In other words, there is a need for a clear a11y reason for any new terms.
Two outcomes of the discussions:
- There don’t seem to be a lot of items people see as compelling to add to ARIA. The community should provide feedback on this (e.g., using the separate dpub-aria repository), and the group can then decide. It may well be that DPUB-ARIA 2.0 is only a few additional features (plus errata management)
- There is a lack of approach for semantic information being added to a publication. This group is not chartered to provide one, but maybe a separate Community Group could be set up to explore that.
As part of the implementation review a first draft of a manifest lifecycle has been created, mostly based on the Readium Web Publication Manifest. That led to a separate discussion on github (see also issues #119 and #122). The lifecycle items was greatly inspired by the corresponding items in the Web Application Manifest (WAM) document, but it turned out to be different from the WAM. At the moment, there are some divergences between these, which must be well documented to back up any decision on the final relationship to the WAM.
A number of other issues are still to be discussed
- how to incorporate the exact browsing context into the lifecycle management
- what happens if a user hits a constituent resource of a WP that does not have a direct link to the manifest
- the lifecycle draft has revealed the relative complexity of the default ordering (based on the WP draft), what to do about that