Joint meeting of the WAI PF WG and the DPUB IG

25 Aug 2014

See also: IRC log


For the PF WG: Janina Sajka, Rich Schwerdtfeger, George Kerscher, Matt Haas, Suzann Keohane, Shane McCarron, Joseph Scheuhammer (clown), Matt King, Bryan Garaventa, Lisa Seeman, Jon Gunderson, Cynthia Shelly
For the DPUB IG: Deborah Kaplan (dkapklan3), Ivan Herman, Luc Audrain, Tzviya Siegman, Paul Belfanti (pBelfanti), David Stroup, Tim Cole, Markus Gylling (mgylling), Dave Cramer (dauwhe), Peter Kreutzberger (pkra)
Ivan Herman
Tim Cole


<trackbot> Date: 25 August 2014

Group references: PF WG, DPUB IG

structural semantics

tzviya: working on building vocabulary we would like to see added in Open Web Platform

tzviya: example: chapter -- concept doesn't exist in HTML

<tzviya> http://www.w3.org/dpub/IG/wiki/StructuralSemantics

tzviya: some of these structural elements have associated behaviors.

a key reason is because of the way some of these elements interact with assistative technologies

Rich move from role to something else.

tzviya -- tag would be managed by PF, but values by IDPF

Rich - you want this to work with HTML Validator

tzviya: Absolutely.

rich - currently limit aria values

Rich -- reluctance to open values up because you might not be able to provide fall back if browser didn't know about values...

clown - some of the APIs take role string as is and stick into a property

<richardschwerdtfeger> <div role=“chapter region”>

Rich - do you see an issue with authors having second value (as fall back)?

markus - may be a non-starter for publishers; hard enough to get them to include basic semantics

tzviya - may not be too hard for larger publishers; but for smaller publishers a decision tree might be too much.

Matt - if native element was <div> you have an element with no semantics

Rich - but not an issue based on current definitions; even if browser doesn't know about it today, assistive technologies have something to work with...

<clown> I'm still getting: This conference is full. Any way to expand the size?

Rich - dpub IG wants to use role attribute, but if browser doesn't know about value, then a second value is needed for fall back

Rich -- that may be a non-starter, but if you have a single value, falling back to native language may be good enough

cynthia - that should be acceptable

jon gunderson -- dom may eventually have a computed role. How would fall back option affect computed role?

Rich - if we had computed role, could compare it to computed role and if they don't match generate a warning

jon 2nd question -- would developers take EPUB3 roles and try to use in HTML 5; would validators distinguish between EPUB3 and HTML5 for validation

Ivan - what is the way a separate group can define new roles while HTML 5 group is doing their thing?

Rich - no mechanism now, but we will have to address long term

Shane - roles accepted by Validator are defined by a document that both PF and HTML 5 can edit

Ivan - do that reference that document?

Shane - they refer to it manually.

Ivan - both groups have write permissions to this document?

Shane - that is my understanding

Ivan - then EPUB could get values added by working with either group, but how big is EPUB list

tzviya about 100 terms

<tzviya> epub 3 structural semantic vocab http://www.idpf.org/epub/vocab/structure/

<dauwhe> Is this the doc in question? http://www.w3.org/1999/xhtml/vocab

Shane - the role attribute spec references the document.

Shane role attribute spec allows for independent role vocabularies to be added by rdfa mechanism

Shane would it be acceptable to have EPUB role values scoped?

markus - some of the EPUB roles will be common, some are more niched (albeit still important)

markus - it could be that the niche roles could be separated out in a prefixed way.

??? if might be a tough call to say which are the most generic, most crucial

<Luc> It would preferable to have a general mechanism for adding vocabulary for EPUB

shaneM - validator was rejecting namespaced role values, but they are not opposed to it, just need to close the loop

Rich - you don't want to namespace everything that is commonly used -- is this a fair assumption

markus - yes

tzviya - the real desire is to create html once and use it everywhere

<Luc> +1

tzviya - we want to avoid namespaces and make it easy to author

<richardschwerdtfeger> role=“epub:chapter”

tzviya - would prefer not to namespace chapter, for example

tzviya - and if you do namespace more obscure roles, they are less likely to get used.

<mgylling> http://www.idpf.org/epub/vocab/structure/

lisa - needs to be successful not just for EPUB authors but also for others with cognitive disabilities

ivan - some of the dpub / EPUB roles are very useful for other communities

<ShaneM> Note that some roles are already esoteric.

<Zakim> janina, you wanted to ask whether we intend these roles only for a11y? Or also for general use?

<ShaneM> "Accessibility is for both meat and machine" is something I get to say all the time.

janina - overarching question: you don't need a disability to appreciate roles; do we anticipate that this will have a general applicability?

several - yes.

Rich - we would do harm if we didn't have a vehicle for opening this up to broader community.

ivan - need to see what is next step. We need to document where we want to go and how.

Rich - we need a joint effort between dp ig and pf wg (aria subteam)

markus - that makes sense. one of the things people in idpf want to know is the size of the effort.

Rich - need to provide that to dp ig but also other stakeholders within idpf

janina and tzviya have an action to start this from the pf and dpub ig sides

ivan - both groups will be meeting f2f at TPAC. we should try to find time to have a session together

Other areas of a11y related collaboration

deborah - what is the why that dp task force can collaborate with PF?

Rich need someone to take lead and run with this.

this will take weekly or near weekly meetings

markus - I think deborah's question was about more broad collaboration

janina - we could write a statement with deliverable as well as with a general statement about liaison

ivan - so what are the reasonable deliverables?

deborah - what are the deliverables?

rich - 1. role attribute value module, 2. a guide for browsers, 3. guidelines for authors

we learned in the ARIA 1.1 effort that we had to write specification for implementation

ivan - still a little in the dark, what would a guide for browser / user agent guidelines mean? what could dp ig contribute?

Rich - each browser maps to accessible technologies on the platform

rich - for a book reader, how do we specify that we want to navigate by book chapter?

rich - does W3 specify or does IPDF?

rich - if you put EPUB semantics in your document, who defines how the browser uses this structure?

markus - that seems like something that could stay (for now at least) with dp ig / IDPF

focus for now on the mappings to accessible APIs

??? how about authoring practices?

markus - there are already some resources on authoring best practices

ivan - dp ig should look at core accessibility guidelines and see how complete they are for dp

deborah - pf wg wants to address gaps as they are identified

george - the semantics that are being introduced go a long way

tzviya - a lot of what deborah is getting at seems overlaps. can deborah work with tzviya and janina

to be clear pf wg wants to help identify gaps, though other groups may work on these issues

ivan - tzviya, janina and deborah will try to have a document by about mid-September. follow-up with call or during TPAC

Summary of Action Items

Action: Tzviya, Janina, and Deborah to have a document by about mid-September.  Follow-up with call or during TPAC

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.138 (CVS log)
$Date: 2014-08-26 07:09:56 $