Publishing Working Group Telco — Minutes

Date: 2018-10-01

See also the Agenda and the IRC Log


Present: Wolfgang Schindler, Luc Audrain, Ivan Herman, Dave Cramer, Jeff Buehler, Wendy Reid, Joshua Pyle, Jun Gamou, Avneesh Singh, Gregorio Pellegrino, Chris Maden, Rachel Comerford, Romain Deltour, Benjamin Young, Ric Wright, Franco Alvarado, Garth Conboy, David Stroup, George Kerscher, Teenya Franklin, Mustapha Lazrek, Charles LaPierre, Vladimir Levantovsky, Ben Dugas, Brady Duga, Lillian Sullam, Ben Schroeter, Derek Jackson, Jean Kaplansky, Reinaldo Ferraz

Regrets: Tzviya Siegman, Hadrien Gardeur, Laurent Le Meur, Deborah Kaplan, Bill Kasdorf, Juan Corona, Marisa DeMeglio


Chair: Garth Conboy

Scribe(s): Dave Cramer


Garth Conboy: shall we approve the minutes?

Resolution #1: last minutes approved

Garth Conboy: minutes approved.

Garth Conboy: Wendy Reid is stepping up as a third co-chair

Wendy Reid: thanks for the warm welcome
… i’m excited to get more involved
… I work at Kobo, for five years now
… I work on both the content and tech side, now I work on the QA team
… I’m also working on audio books

1. Audiobooks

Wendy Reid: the audio books task force has met twice now
… the first conversation was to define audiobooks
… which was contentious
… see Hadrien’s example in issue #247
… an audio book is audio-only primarily, has a toc, duration, author and reader, etc.
… it’s not SMIL, not media overlays, not video
… but there’s lots of commonality there
… that’s where we are right now. We have a few issues we’d like to discuss with the group
… as these issues also affect other types of web publications

George Kerscher: this specification seems to be very much needed now
… I’m wondering if the implementations of this as it is being worked on is… is going to be leading… will this be implemented before the rest of the spec is approved?

Wendy Reid: that’s a good point
… tzviya reached out to BISG to talk with someone from the audio bookseller association
… but we need buy-in first. We need to make friends.
… We need industry buy-in before we make a spec.

George Kerscher: I’ve visited with them in the past about audio book specifications. They do need help. It’s not a technical group.
… they’ll need a lot of help and tools.

Jean Kaplansky: present

Garth Conboy: audiobooks are particularly interesting… for other things there are lots of existing standards
… but there is no normal way in the audiobook world
… so it could really help, and help validate the whole spec

George Kerscher: the DAISY standard has not been adopted, as it was too complex

Wendy Reid: could you share your contacts with us, so we can talk to them?

George Kerscher: my contacts are old, and I think management has changed.
… going through BISG for the USA sounds like a good start.

Avneesh Singh: we don’t have many readium people, but readium has plans for implementing audio books. I don’t know the timeline.
… about the definition of audio books, this is the definition of minimum viable audiobooks
… there is a lot of scope for building a more sophisticated audiobook

Ivan Herman: for my understanding from the current draft
… when you speak of audio books, what you mainly make use of from our standard
… is that you can use any resource for the web, so the book is a bunch of audio files
… but we have an entry page in html
… does that create problems?
… there are also additional attributes to add to the manifest, which is OK
… is there some more complicated technical issues that have to be solved?

Wendy Reid: yes, if you look at Hadrien’s sample file which works in a web browser
… . there’s a manual navigation process
… I don’t see a problem with an html entry page
… the challenge for user agents will be to interpret the reading/playing order to make it seamless
… our sample right now is not seamless

Ric Wright: Yes, Readium is developing audio support, but you should talk to Hadrien

Wendy Reid: let’s discuss the audiobook issues
… 307 and 308, about duration and bitrate

1.1. duration, ISO format

Ivan Herman: link: Issue 307

Wendy Reid: one of the challenges is the relationship between track length and chapter length
… so how do we express duration in an audiobook?
… schema has audio duration, using an iso format (ISO 8601)
… this could also affect video content
… our alternative would be to extend publicationLink
… we also discussed bitrate and format, which would be expressed similarly
… this also drifts closer to best practices
… do we want to recommend bit rates and formats, or do we want to leave it to the market?

Ivan Herman: I don’t fully understand the options
… there is the matter of the duration property, and what it’s value can be
… I’m not fond of the ISO format
… but I still use them
… I would be hesitant to invent our own format to replace the ISO one
… we shouldn’t do that
… we should just deal with the ISO format
… this is also the format recommended by
… so we should follow what they do
… we can raise the issue with at TPAC, and danbri will be at one of our sessions
… that is also true if we take the duration as a property for publicationLinks
… even if we have defined the object, we still need to be aligned with
… so no matter where we use duration, we should keep to what’s defined in ISO

Benjamin Young:

Benjamin Young: html5’s time tag and datetime attribute all use ISO 8601, so it’s the prevailing winner
… even if you find it ugly

Ivan Herman: we do

Wendy Reid: no one is arguing, but we needed to bring it up to the larger group
… schema has a value for audiobooks, but it might be harder to extend to video

Ivan Herman: that’s a different question

1.2. bitrate and format

Ivan Herman: link: Issue 323

Wendy Reid: the other issue is bitrate and format
… there’s a secondary issue, #323, logged by Avneesh
… there are innumerable ways to encode audio…mp3, flac, opus, etc
… and many different bitrates
… in EPUB 3 we had the idea of core media types
… with WP, do we want to express recommended media types, more best-practices?

Ivan Herman: what does HTML say about the audio element, in terms of bitrate and format?
… we should should push the corpse into someone else’s courtyard

Avneesh Singh: audio element in HTML accepts a broad set of formats
… but popular browsers may use a smaller set
… we should not restrict WP in the way we restricted EPUB
… we should not have core media types
… but we do need to recommend formats for audio publishers which will actually work
… the only question is how to present this information? It could be in a non-normative document
… note I’m not mentioning EPUB4 here

George Kerscher: there are issues where human audio is in a smaller bandwidth than music
… so some formats are better for human audio
… and variable bitrates can make it more difficult to seek to a particular place
… and the available of time scale modification, to speed up or slow down without changing pitch

Wendy Reid: by identifiying use cases maybe we can find some recommendations

Benjamin Young: this is HTML’s section on mime types, which is more generous than what you’d find in EPUB
… this also allows for innovation and progress, because it is not restrictive
… let’s keep the web in web publications :)

Joshua Pyle: looking at George’s use cases for seeking and time scale modification… do we have these use cases? Should we add them to the general use cases doc?

Wendy Reid: we should have audio-specific use cases

Ivan Herman: one more source to look at…
… there are two related groups at W3C
… there’s the audio working group
… but what might be more interesting is the media and entertainment interest group
… it grew out of webtv but went beyond that
… at the Tokyo workshop the team contact was there, and I tried to convince him to join our F2F meeting
… so we could talk to him there
… I think we already discussed that we would not have a restriction on web publications–whatever works on the web works for us

Wendy Reid: some of these issues could be cleared up by use cases. We’ll discuss in the task force.

1.3. reading order

Ivan Herman: link: Issue 322

Wendy Reid: the last issue is #322
… it came out of #320 on mixed media and the reading order
… some audio books may have supplemental visual content
… often added as PDF or EPUB, but no UAs support this
… could a WP support this functionality? Should we have mixed media in reading order?

Ivan Herman: from the WP point of view, more difficult to restrict it than to have it
… we don’t have a restriction on the reading order–it can be any resource. PDF or drawing or whatever.

Garth Conboy: your statement is correct
… what happened with #320? Why is it ready to close?

Wendy Reid: because 322 addresses the question better

Garth Conboy: getting a PDF with a bunch of MP3 files is common
… so we’re in a good position with what we have
… the reading order would be a bunch of mp3s.
… then the other stuff would be in the “resources” as it’s supplemental
… . out of band auxiliary resources go in resources
… not sure about having other stuff in the reading order

Avneesh Singh: for 322, what garth said is right
… they were discussing the practicalities–what if the reading system doesn’t support PDF?
… what should reading system do if the media isn’t supported?
… Hadrien mentioned a profile for audio books, where the reading order might be audio-only

Brady Duga: avneesh hit on my issue
… should we say what happens when these things happen in the reading order
… I like allowing these things
… but I don’t know how far we should go in the spec to require certain behaviors
… I don’t want to prevent innovation

Joshua Pyle: I agree with that
… to avneesh’s point on what if a reading system doesn’t support PDF
… a web publication should support anything the web supports
… but what if Alexa comes across a PDF? What happens?
… we can’t say what every implementation will do

Ivan Herman: back to one comment by Avneesh about profiles
… in a sense I could say we already have them, just not by name
… we require the type of publication in the manifest
… a reading system looking at the type can find out that it is an audiobook
… we already say that certain schema terms only make sense in certain contexts
… no need to make things more confusing

Benjamin Young: I’d agree with ivan
… I’d also like to mention atom and rss, which started as text-only but then podcasting happened, and now both formats represent both audio and text, and possibly intermixed
… and you’d just listen to audio if your device has no screen
… but on some other device you could consume both types of content
… so we don’t need to overly constrain the world
… we don’t want the types to constrain things.

George Kerscher: I agree with the ideas of not limiting innovation
… with AR, VR, etc
… however, if a reading system ignores what it doesn’t understand, the audio book would be understandable if you just consumed the audio portions

Wendy Reid: that’s the assumption that most user agents make with audiobooks
… we just don’t offer the supplemental content, for example
… we don’t want to stifle innovation. Supplemental content is important.

1.4. further steps

Wendy Reid: the task force has work to do, including creating use cases

Garth Conboy: as far as resolving 322, continue on the issue, maybe discuss at TPAC, but keep working on issues in the task force

Wendy Reid: we don’t want to prevent different content from being in the reading order
… maybe use cases can provide clarity
… the task force can deal with this, and come back to the group at tpac

Ivan Herman: trying to organize ourselves…
… could we have a proposal of what features we’d add to the WP draft for Audio?
… after TPAC we could publish a new draft, that would have a solid basis for public comment on audio?

2. TPAC agenda

Ivan Herman: link:

Garth Conboy: we have a pointer to the TPAC agenda, and it’s open for commenting
… it’s not yet cast in stone
… I’m out next week in Frankfurt, but Tzviya is back
… any other business?

Brady Duga: there’s a Monday night group dinner in Toronto. That might be complicated logistically ;-)

Ivan Herman: Oops, cut and paste error :-)

Garth Conboy: thanks wendy and thanks everyone

3. Resolutions