Publishing F2F, 2nd day — Minutes
Date: 2018-10-23
See also the Agenda and the IRC Log
Attendees
Present: Joshua Pyle, Leonard Rosenthol, Juan Corona, Garth Conboy, Ivan Herman, Avneesh Singh, Ralph Swick, Karen Myers, Bobby Tung, Wolfgang Schindler, Charles LaPierre, Wendy Reid, Gregorio Pellegrino, George Kerscher, Rachel Comerford, Dave Cramer, Toshiaki Koike, Romain Deltour, Brady Duga, Luc Audrain, Reinaldo Ferraz, Liisa McCloy-Kelley, Jun Gamou, Marisa DeMeglio, Tzviya Siegman, Hadrien Gardeur, Benjamin Young, Laurent Le Meur, Daniel Weck
Regrets:
Guests: Kazuyuki Ashimura, Judy Brewer, Jeff Jaffe, Judy, Naomi Yoshizawa, Guillaume Sire, Janina Sajka, Lloyd Rasmussen
Chair: Garth Conboy, Tzviya Siegman, Wendy Reid
Scribe(s): Karen Myers, Dave Cramer, Romain Deltour, George Kerscher, Daniel Weck, Brady Duga, Garth Conboy, Juan Corona, Laurent Le Meur
Content:
- 1. Joint BG/WG meeting
- 2. Synced media
- 3. EPUB4
- 4. Next F2F and other meetings’ schedule proposal
- 5. Audiobooks
- 6. Resolutions
Ralph Swick: -> https://www.w3.org/2018/10/22-pwg-minutes.html previous minutes 22 Oct
Ivan Herman: https://www.w3.org/publishing/groups/publ-wg/Meetings/Minutes/2018/2018-10-22-pwg.html
1. Joint BG/WG meeting
Ivan Herman: (This session is chaired by Liisa McCloy-Kelley and Luc Audrain, BG chairs)
1.1. what does the WG want from the BG on priorities? What about EPUB 3.2?
Liisa McCloy-Kelley: let’s talk about what direction the WG wants from the BG
Garth Conboy: from yesterday, we spent the whole day focusing on where the group has been heading with WP
… a methodology to define a publications and its bounds so it can live on the web
… that seems interesting to the education and journals space
… less interesting to trade publishers
… we’ve just “finished” EPUB 3.2
… which is significant for the classic EPUB market, especially as epubcheck gets updated
… I think it’s helpful to have the latest version of EPUB actually being used
… I’m concerned about EPUB 4 stepping on the tail of EPUB 3.2
… I don’t want to derail EPUB 3.2
… perhaps by focusing on audio books
… what do you think of priorities? 3.2 vs 4
… this WG operates independently by W3C process
… but the WG can take input from the business group and the publishing community
Gregorio Pellegrino: for priorities…
… we need some clarifications about web publications, portable WP, and EPUB 4
… what are the differences?
… we have to explain in a clear way what will be the differences
… in Italy we had a big problem moving from EPUB 2 to EPUB 3
… so it might be harder to move to EPUB 4 or WP
… we want an infographic that’s simple and easy to understand
Romain Deltour: garth said the question of conflict between 3.2 and 4
… the conflict already happened between EPUB 3.2 and WP
… I don’t see the business case for web publications
… we want to be webby, but the browsers aren’t involved
… I don’t think the publishing industry won’t migrate easily and quickly
… we say that if WPUB is adopted, it will attract interest from browsers
… but I don’t see forces that will lead to broad adoption
Benjamin Young: I work for a publisher. Ww have a business case for WPUB, but not for the existing spec
… we don’t use EPUB for journals
… I have not understood how WPUB ended up being part of the EPUB lineage
… I want to see WP shaped by web publishers
… but we don’t want to have the constraints of EPUB
… I want to split these things–EPUB and WPUB on separate paths
Garth Conboy: I view WP as not so much fighting with EPUB 3.2, as they are different animals
… one on web, one packaged
… those can progress at the same time without conflict
… WP getting traction, and moving core content to be OWP technologies, and figuring out how to package that… seems to work
… I view “putting books in EPUB 4” would be solving a problem that doesn’t exist yet
Rachel Comerford: if I understand what you’re saying, I disagree
… going into WPUB, we’re trying to solve a problem we haven’t defined yet. That’s our biggest challenge as a WG
… we declared a solution to a problem we didn’t define
… so now we’re stuck where we resolve problems two ways–at the high level, by not looking at the details
… or we get stuck bike shedding, and don’t see the forest for the trees
… it doesn’t feel like we’re getting closer to an implementable specification
… it’s been frustrating for me
… we have so many tools available to us in terms of progress and product development, but we’re not using them
… and there’s a lot of infighting over individual solutions
Garth Conboy: I don’t think we disagree
… I don’t view WP as in the same space in EPUB
… it’s a different problem set
Tzviya Siegman: focusing on question of priorities for BG to WG
… I very much agree with Rachel
… we have lost our way, and getting feedback from the BG is good
… we started off with a vision of WPUB
… but we haven’t gotten to a great place with that
… if the BG wants us to focus on EPUB, maybe could do that
… right now Wiley produces articles as HTML
… so things of interest to us are
… AMP, packaging, identifiers, addressability
… annotators, selectors, web components
… we haven’t solved these problems, and our spec doesn’t address them
… maybe we want to focus on EPUB in the WG
Avneesh Singh: touching on many things that have been discussed
… we should market the best product for each segment
… we are trying to keep the products near each other
… the benefit the consumer is to use the browser
… so trying to keep things close to EPUB does not help
Joshua Pyle: +1 to Avneesh
Avneesh Singh: we should have WPUB go along the web direction, working in the browser
… and for the publishing industry, follow the EPUB path
… there’s no benefit to keep these two things similar
Benjamin Young: +1 to Avneesh
Romain Deltour: +1
Luc Audrain: I would like to emphasize that the work on EPUB is important for us
… especially as EPUB 3.2 now has a better relationship with HTML and CSS, using the latest versions
… we have momentum now to help us to start the traditional publishing industry to produce good modern content
… our problem today is that HTML and CSS doesn’t work everywhere
… we want to improve the reading experience for a11y, layout, etc
… EPUB 3.2 is the best EPUB ever
… I understand WP is important for journals
… I’m concerned, like Garth, that we might disturb the momentum of EPUB 3.2 with EPUB 4
… except for audio books, which we could prototype EPUB 4 for the market
Ralph Swick: I’ve talked to some of you
… I’m hearing a new message that w3c didn’t appreciate two years ago
… our original proposal was to make epub a native part of the web, viewing them in web browsers
… we had a few ideas
… some starting points
… but we didn’t survey every touchpoint
… we proposed to start the WG
… what we didn’t fully appreciate was the degree to which the industry was moving from EPUB 2 to EPUB 3
… EPUB 3.2 is a significant improvement, as it updates the relationship to HTML and CSS
… that does remove some of the message about what the new working group objectives were
… it reemphasizes the need to say EPUB 3.2 is a big step ahead, and a step on the road to more integration
… I see that the BG can help make people know that this a step to the future
… this is an intermediate step
… I’ve heard different emphasizes… the spec is good enough and we should market it more, for example
… and a suggestion that there’s more work to do around interop
… what are the next steps you want to happen with 3.2
… and are there resources to do any of this work?
… we need to be careful about the messaging around both these things
… we’d like a clear message on what you want the next steps for 3.2 to be
… as well as working on the future
… we need guidance from the BG on how we balance these things
Hadrien Gardeur: there’s no solution for audio books right now
… we get bad stuff from publishers
… both EPUB 4 and WP are really relevant for audio books
… EPUB 4 is relevant as an ingestion format
… WP is relevant as a delivery format, sent to users
… the same thing is true for comics
… those publications can be huge
… we don’t want to send 100s of MB to a user
… the ability to stream resources is incredibly interesting
… there are massive markets around audio and comics
Brady Duga: I agree with Rachel’s summary of the WG issues
… specs work best when two or more groups come together to make something interoperable
… when people just blue-sky things it doesn’t work
… it’s not about the quality of the spec
… you can write a crappy spec like EPUB 2, but it actually solves a problem
… audio books is a place where we have a problem, and there’s not a good solution
… it’s interesting we split EPUB4 and WP
… there is a difference with streamed vs packaged audio
George Kerscher: I like WP, the concept
… the fundamental is that web browsers are the most accessible interface
… and for journals that are one document, I can see it working
… we have a disaster on the web right now with publications being out there
… for example WHO (world health) has lots of information that’s completely inaccessible
… getting those moved to WP would be a great service to society in general
Tzviya Siegman: if we shift our focus to EPUB
… the work on audio books could be standalone
Luc Audrain: re: audio, WP and EPUB 4 are important as Hadrien said
… I would suggest we don’t call this EPUB 4,
… we should have audiobooks
… this message would focus on something we need now: AudioPub
… with EPUB 3 we have the best with HTML and CSS
… I don’t know if we have to shift the wg for this
… I think we should keep momentum on EPUB 3.2, to reassure the industry
… but use audio books question to promote WP, but call it AudioPub
Benjamin Young: I think we modeled the group upside down
… we created specs from whole cloth, which should have been done in CG
… and EPUB should be done in a WG
… and move audio and web pubs to CG to explore and experiment
… and then make specs out of the overlap, and then come back to the WG
… that we put the specs in a place where exploration is encouraged, and experimentation is fine
… then come back to the WG
… the actual spec is 3.2, the experiment is WPUB, the dream is EPUB 4
Tzviya Siegman: Publishing CG http://www.w3.org/community/publishingcg/
Garth Conboy: I agree with all y’all
… I’m sensitive to the 3.2 is important, we don’t want to do EPUB 4 for ebooks right now
… we could use AudioPub as a way to learn about the future of WPUB
… and we could work on comics
Charles LaPierre: +1 well said BY
Garth Conboy: that could be done now but not dilute 3.2 and not confuse the market
Avneesh Singh: re: segmentation
… this group should sustain WP in more webby way
… to have WP working in browser–this is one segment
… the confusion is EPUB 3 vs EPUB 4
… maybe continue with EPUB 3
… and using a profile for AudioPUB with new tech
… traditional publishers continue to use EPUB 3.2
… and audio books which is direct outcome of new work
… 1. WP for web
… 2. EPUB 3.2 for traditional publishing
… 3. AudioPub as a profile of WP
… .those are the three segments
Gregorio Pellegrino: +1 for Avneesh
Ivan Herman: my problem is, regardless of what we do where…
… whether we follow Benjaminjamin’s model
… or continue this group
… our problem is limited resources to do the technical work
… we could reach out and say we’ll do 3.2 and WP at the same time
… but we don’t have the people to do the technical work
… 3.2 will require a lot of work
… the spec work is done, but we need to do test cases, figure out interop
… we lack the personpower we need
… if we can’t get new members, new engineers that can do the work, then all of this is just a dream
Benjamin Young: +1 to ivan’s concerns
Marisa DeMeglio: I hear lots about traditional publishers vs web publishers
… why are these separate islands
Tzviya Siegman: traditional in this group means trade
Marisa DeMeglio: trade publishers don’t want to be on the web
Marisa DeMeglio: I find that WP, PWP, EPUB4, etc to be really confusing
Laurent Le Meur: we could remove one, PWP
Charles LaPierre: when we released 3.1, the problem was we didn’t have epubcheck
… I know DAISY is working on 3.2 epubcheck
Tzviya Siegman: https://www.w3.org/blog/2018/10/epubcheck-fundraising/
Tzviya Siegman: WE NEED DONATIONS. GIVE US ALL YOUR MONEY NOW.
Garth Conboy: epubcheck is getting the focus it deserves
Benjamin Young: re ivan’s concern about peoplepower
… by using a CG you open the doors to engineers to participate in a more familiar environment
… there’s not a membership obligation
… early-stage startups aren’t in a position to be members, do calls
Juan Corona: +1 bigbluehat
Benjamin Young: but they can write scripts to prove or disprove a point
… and perhaps if their work pans out, get in a position to join w3c
… CG works for small operators
… when there’s some momentum, larger companies may become interested
… to marisa’s point
… pay attention to verifiable claims, identity stuff, all that kind of thing
Brady Duga: re: ivan’s comment on developers
… developers are not interested in implementing WP
… I’m not interested because I sell ebooks
… the technical details are uninteresting to me
… but in the EPUB 3, I was involved and these things mattered to me
… I attend the audiobooks calls, because it matters to me and what I’m doing
… perhaps the lack of interest in implementation is keeping people away
Ivan Herman: which should mean more members
… it’s easy for you because you’re a member
… the problem is twofold–both membership, and getting engineers from members
Avneesh Singh: Maybe we are trying to solve all the problems together
… good to separate; one bucket is WP
… have CG kind of thing
… one practical thing, this WG is more than half way through
… rolling everything back is a huge overhead; maybe we move innovation along
… other bucket is EPUB3.2
… to Ivan’s point on manpower
… maybe we don’t rush to put EPUB3.2 onto the rec track; see what industry does
… maybe look at how market reacts
… Third segment is audio books; we can work soon and start pushing
George Kerscher: would 3.2 on rec track change publishers’ perspective
Garth Conboy: I think coming out of CG, in my view is seen as an established work
Wendy Reid: When I go to my engineers, this is latest update
… there is no work happening until we get publishers interested
… cannot get people interested until I get publishers ready to support
Liisa McCloy-Kelley: if we support EPUB3.2, I’ll ask you what you need to support it
Wendy Reid: We are happy to take that
Benjamin Young: we buy scholarly
… loads of vendors at ebookcraft, Toronto, March
… most won’t be in WGs but would like to have some say
… and do development to be future of what is being done; would like to see their involvement in a CG
Leonard Rosenthol: I would like to comment on Ivan’s comment
… my perspective on this
… this group and the work that this group has been doing is excellent
… a lot of great potential with WP and audio and the like
… this group is 90% publishers, or people from publishing industry coming from IDPF
… that is great; we have leverage that experience to build the work we have done
… the reason you are not seeing other folks come in, that’s because it’s that type of focus
… You have heard me say, my focus is that Web Publications encompass more
… work we have done helps to accomplish that
… while we have some of the other things in this group
… it does give folks that opinion outside of here and other standards bodies meetings
… they just don’t here a place for them here
… Making WP a separate piece of work from the ebook work is good
… will help drive toward Ivan’s desire to see new blood come in
… provide a place to build separate from the ebook community
Liisa McCloy-Kelley: summarize what I think you all said
… Everyone agrees 3.2 is good, it’s out there
… question is whether EPUB3.2 as good as it is, or do we need a rec track
… and does that mean anything to the BG or a bigger group of people
… figure out if that should be prioritized
… Flip things for CG to incubate thing
… then move to BG for input, then move to WG to work on
… supported pretty good
… Then there is a question of Audio and whether that is a priority
… and if WP is a priority
… should they run simultaneously
… requirements…
… does that sound right?
Wendy Reid: We don’t have feedback from the industry
… just Brady and my experience
… Would love to know what audio industry wants; what problem we are trying to solve
Liisa McCloy-Kelley: and what will it take to implement and get broad support across reading systems and from publishers worldwide
… One thing somebody brought up, not sure where it fits
… idea that comics and audio could both work in same time frame, or same chunk of the puzzle
Luc Audrain: that is a good transition
Luc Audrain: I think about Tokyo workshop
Garth Conboy: shall we close the queue?
Ivan Herman: I think it’s important to realize the BG
… putting 3.2 on rec track
… personally I think it would improve the spec
… by creating the test suites and test environment
… which leads to adjusting text to make it more precise
… at end of process
… we come out at W3C will be of a higher quality
Karen Myers: ..and will lead to better interop and compliance with reading systems
Ivan Herman: a bit technical, that should be part of the decisions
… not just putting standards stamp on 3.2
… downside there is still work to do
… I think this is work that would be highly beneficial
Liisa McCloy-Kelley: What do we mean by compliant reading systems and how would you test for that
… when majority of reading systems don’t @
Ivan Herman: good question; don’t have answer there
… but goal is not to have to test reading systems on 15 reading systems
… try to avoid the “best viewed on” issue
George Kerscher: companies have been ingesting from 3.2
Ivan Herman: I cannot answer details, this is part of job
Luc Audrain: 3.2 is not a web publication
… I am concerned
… 3.2 is ok
… for publishing and not rewrite
Ivan Herman: not rewrite
… 3.2; maybe some minor details
… I am not saying at all to change it
… speaking here with guidance of [Dave C]
… no testing environment; rec track would force us to do that
Laurent Le Meur: rec track maybe would include that
Ivan Herman: It’s not implemented; why not there
Karen Myers: Liisa; We can dream of a future that includes JScript
Karen Myers: Ivan; what implementation means
Ivan Herman: if MathML is in MathJax…
… whole thing is useless if not implemented
Jeff Jaffe: If rec is that the WG picks up 3.2 or Audio Pub
… we are not chartered today
… we would need to recharter
Tzviya Siegman: Was going to say what Jeff said
… turn to Dave to speak about 3.2
Dave Cramer: I feel like we are confusing means and ends
… 3.2 is not our goal
… our fundamental problem with EPUB today is interoperability
… EPUB3.2 is not significantly different from 3.0.1
… tech changes are quite small
… except for core of HTML and CSS
… in some context we got that for free before
… a lot of reading systems are based on browser engines
… not like reading systems will unblock certain features from happening; they never filtered on it anyway
… not changing what’s possible in EPUB
… if 3.2 goes on rec track
… what is point of making a first spec
… go over a hundred times
… but learn from experience of doing tests
… why go down that path
… make this more interoperable
… what does reading system X or Y do or not do
… or unimplementable because it was not tested
… use this as a tool rather than an end in itself
1.2. Tokyo workshop and its results
Ivan Herman: Report of the workshop: https://www.w3.org/publishing/events/tokyo18-workshop/report.html
Luc Audrain: reflowable issues
… also books with complex layouts
… too sorts: graphic arts; manga, bandes dessinées
… and those with complex layout
… today EPUB is solving with fixed layout EPUB
… layout is preserved using open web technologies
… for manga, it’s an image of the page, fixed by a capture by the page in print PDF
… for other books it uses absolute positioning with page as viewpoint
… put text and image on the page
… it breaks solutions
… put on market mangas and fixed layout text for text books or cookbooks
… there is a market for it
… market is very strange
… for fixed market there is 50% market in Japan for manga
… if we change things here, we have to pay attention
… also success in Korea using web technologies
… these two approaches were exposed in Tokyo workshop
… two ways to solve this
… Idea is to have a better issue in digital books for this fixed layout
… for manga and bandes dessinées; understand what is expected by the author
… describe page with a specific taxonomy
… some work was exposed
… some thought it was good and wanted to join this work
… also a W3C CG was created
… for international work on this
… how things are expressed by the author in this taxonomy
… bring two together in future
… CG is one outcome
… second kind of group
… make more accessible text books, express as author wanted as a page
… also be better on small screens
… we had good presentations on capacities of CSS
… there could be a layout system for HTML
… it was interesting for trade publishers
… use InDesign
… lay out important, meaningful layouts
… layout means something
… some people working on this
… to try to figure out different part of the layout and make it available in a reflowable manner
… at end of day
… we had reflection about two ways to make this complex layout more reflowable
… one as Dave explained is to compose on HTML and CSS
… with CSS we may have books created with complex layout directly
… means a lot of work to be done in web browsers where layout is built
… not as sophisticated as what we can do now
… Other way is to express perfect job
… in complex layout in HTML and CSS; some kind of conversion
… need to understand
… could be expressed as dutifully as an InDesign page
… and on smaller pages, more free
… the two directions we identified
… We need more input from publishers in CSS WG
… and what has been specified needs to be implemented in browsers
… First, on long long-term
… it would help us to go a step further than fixed layout
… help us to do some important business in manga
… Go to another step in the future when we have this for complex layout in HTML and CSS
… we hope the whole publishing industry will contribute to this evolution of CSS
… Others who attended Tokyo workshop, anything to add?
Ivan Herman: one more thing that did come up
… which is not necessary a job for this group
… evolution of CSS is
… for many things, how we use CSS for various layouts, much based on frankly hacks
… old days we used tables
… then told not to do that
… situation where we throw out everything from CSS an start over again
… flexbox and grid and combination of the two create a new rich, clean environment
… but more difficult for those who have to carry the past
… would need a lot of education and outreach
… try to get this kind of outreach for users might be important
… I realized I may have to relearn things
Laurent Le Meur: There is a lack of tools coming from a fixed page from InDesign
… to interactive
… also importance to lack of EPUB3 for lack of handling manga
… but complained as images
… they don’t have @ they would like
… segueway to what we would want manga group to do
… not move to 3.3. but use W3C techniques and implement another way
Leonard Rosenthol: I agree, Laurent, there is a lack of tools because it’s really, really hard
… not a shock to say we have invested hundreds or thousands of dev manhours to solve that problem
… nobody likes what we have come up
… with; when it comes to responsive design
… you cannot guess things
… you either put in manual controls; put in break points and the like
… that is too complex for public, although matches CSS model
… or we go to heuristics and machine learning
… great for average user but developer complains
… it is a really hard problem
… if you want response design, you have to start with responsive design in mind
… that is very difficult
… what does authoring in a responsive environment look like? That is a difficult problem to solve
… we agree
… and why CSS WG chair works for Adobe
Laurent Le Meur: 90 percent of designers are not developers
Leonard Rosenthol: right but those designers turn their work over to developers
… people want to add these same things into their web sites
… how do you produce content that delivers all those results
… as we look at it
… just putting it into context
Ralph Swick: [to Leonard’s point, one of the many things I want publishers to teach the Web is how to bring the centuries of knowledge in book design to web tooling]
Ivan Herman: one more thing came up in the workshop that has a general message
… there has always been a strong reluctance
… to consider using JavaScript for anything in the Publishing world
… there are accessibility and security issues
… I think that is a general reaction to look at a again and reconsider
… we have to accept that layout effects are done in JS
… some examples in the workshop, Rachel @
… we know about Houdini project in CSS work
… that gives you environment to plug in CSS
… have to incorporate in our thoughts
… publishing world may need to do active scripting
… active push away from scripting is not tenable for long
Luc Audrain: https://www.w3.org/publishing/events/tokyo18-workshop/report.html
Luc Audrain: Add on irc
… one to the report that the chairs have written
Luc Audrain: https://www.edrlab.org/2018/09/26/w3c-tokyo-workshop-part-1/
Luc Audrain: and the second is an EDRLab report on this subject
… Laurent has three-part report
… links to all the presentations are in the full report
… way to have messages from this
Garth Conboy: time for a break
… sounds like you have some homework, Liisa
… you will go back to BG and come back to us
Liisa McCloy-Kelley: thank you everyone
Ivan Herman: —- BREAK —-
2. Synced media
Marisa DeMeglio: let’s talk about sync media and publications CG
… what our findings are, what are the next step
… what’s the technology selection process
… we want a way to sync pre-recorded audio clips with text in a web publication
… we don’t want to interfere with the Web Pub itself, but have an overlay with audio clips
… at last TPAC we showed what that could look like
… so you have the text of your book displayed, some chunks highlighted, and at the same time you hear the audio for that chunk
… the playback follows the highlight
… in order to do this we looked at the preexisting tech
Marisa DeMeglio: https://github.com/w3c/sync-media-pub
Daniel Weck: Community Group: https://www.w3.org/community/sync-media-pub/
Marisa DeMeglio: https://github.com/w3c/sync-media-pub/blob/master/technology-selection.md
Marisa DeMeglio: it started at TPAC last year, and we looked at what was out there
… the 1st thing we considered is SMIL, which is used successfully in EPUB 3
… it can contain a pointer to the text document and a pointer to the audio clip
… unfortunately it’s not very active as a standard, and has no browser support
Liisa McCloy-Kelley: it works for picture books
Marisa DeMeglio: yes, but it’s not the most developer friendly format
… next is TTML2, but there’s no way to point to a text HTML document
Dave Cramer: TTML = Timed Text Markup Language https://w3c.github.io/ttml2/index.html
Marisa DeMeglio: you could hack it, but it’s not convenient
… Web VTT does have browser support
… but then you can’t have nested, like you got in Media Overlays
… so you can’t have different granularity (read by paragraphs, then read by words)
… Web Animations, is a cool spec but has no declarative syntax
… WebAnnotations has the right degree of pointers (text and audio) but has no processing model (can’t play it from start to end)
… recently we were pointed at multidevice support and timingsrc
… but no processing model either
… next thing we want to look at is to roll our own
Leonard Rosenthol: did the group evaluate the possibility of going to TTML or WebVTT WG and asking for the necessary changes?
Marisa DeMeglio: yes, I spoke to them
… having the text format in TTML just doesn’t work for us
… ad they suggested to transform the content on the fly to HTML, but it’s not realistic
Daniel Weck: TTML is more advanced, it has its own layout and rendering model
Ivan Herman: what was the goal of TTML?
Daniel Weck: settop boxes, captions
Daniel Weck: https://github.com/w3c/sync-media-pub/blob/master/manifest-exts.md
Daniel Weck: we’re gonna look at the various additions we can make about the manifest
… and examples of JSON serialization for the model we’re proposing
… not deviating much from EPUB 3 MO
… but conceptual extensions
Romain Deltour: [looking at the doc linked in IRC]
Daniel Weck: we now that schema.org defines an audio book type, as an extension of book
… how we deal with Web Pub manifest when we want to enable a union of features from the book type and audiobook type
… do we want to explicitly refer to both type?
Ivan Herman: no
Daniel Weck: we want to rely on the audiobook type as it gives us additional properties
… one of them is the ‘readBy’ property
… I’m skipping intentionally the discussion about the video book type
… we’ve discussed the syntax for the values of the duration
… s/duration/duration property/
… we propose various syntaxes
… we can align with the Media Fragment syntax
Ivan Herman: I’m wondering if the discussion we had with danbri is related to this
… I don’t think schema.org would want to exclude the ISO syntax, but we could also extend to a simplified syntax
Daniel Weck: if we want RS to support the spec, they’d have to implement support for these different syntaxes
… the onus is on the RS developers, we have to be mindful of that
… I’m not sure what it means in terms of process
Tzviya Siegman: we can comment and follow up
Daniel Weck: the duration property applies to the whole publication
… but can also be applied to individual links in the reading order
… in the case of a regular publication, audio files can appear in the resources and the duration property can apply there too
… these links are currently of type publicationLink
… but we could use the schema.org link role to make it more extensible
… about the properties specific to sync media
… in EPUB 3, when the text fragment is actively narrated, we set a color highlight with a CSS class
… there’s a default CSS class and RS which support MO will inject that class in the doc
… since then, the CSS Selectors Level 4 introduced a new pseudo class called ‘:current’
… and two other called ‘:passed’ and ‘:future’
… the stated use case is TTS read aloud, but it doesn’t exclude our use case
… we don’t know what’s the state of this pseudo-class in the CSS WG / specification process
Ivan Herman: how do they define what “current” means?
Daniel Weck: to me it’s still very much unclear, let’s keep an eye on what’s going on there
Dave Cramer: https://drafts.csswg.org/selectors-4/#active-pseudo
Ivan Herman: eventually they will have to define some sort of processing model and we’ll have to see if it works for us
Daniel Weck: correct
Ivan Herman: shouldn’t we talk to them right now when it’s still very open?
Daniel Weck: yes, we just discovered this very recently
Tzviya Siegman: is it the same thing as aria-active?
Daniel Weck: it intersects
… what we’re trying to achieve is to have enough in the mainstream manifest so we can use it for sync media
… we need a number of hooks in the Web Pub manifest format
… one of the issues we’ll have is whether we want to group them under a namespace?
… important: we need to associate an HTML doc in the reading order (or resource list) with a synced media overlay
… we want to reference the overlay with a link
… we’re proposing an additional property called “sync-media” which is a nested publication link
… it’s implemented in Readium 2 (with a slightly different syntax)
… so we now it’s implementable
… that link would point to something equivalent to EPUB 3 MO, but simplified and using a JSON syntax
… we propose a media type for that JSON format “application/vnd.wp-sync-media+json”
… registration to IANA to be discussed
… it offers RS a way to discover the synced media overlays in the manifest, rather than checking for the extension or file content
… we discussed whether the duration should be attached to the main text document or to the JSON overlay
… we propose to attach it to the overlay
… Issue 6: one proposal is to have synced media resources in addition to resources and reading order
… having them grouped for ease of processing
George Kerscher: so if I understand you have a regular Web Publication, and an audio publication, then a synced media that connects them both?
Marisa DeMeglio: I think that references one of our overarching principle
… MO should be able to be packaged separately
Daniel Weck: and you don’t mean a ZIP file, but a way to define the resources together
Marisa DeMeglio: yes, and we want to be compatible withe Audio Book TF
Brady Duga: does this proposal still requires IDs on all elements that we want to highlight?
Marisa DeMeglio: great question, we have to figure this out
Daniel Weck: the grouping of resources is not specific to synced media, it can be a good use case for mainstream publications as well
… it can organize a very long list of resources
Ivan Herman: I understand the intention, but as we defined yesterday the boundaries of the Pub, we’d have to add up all these resources to the definition
… I’m a little bit worried about this kind of effect
Daniel Weck: the meat of this proposal is a new format, the JSON media overlay
Marisa DeMeglio: https://github.com/w3c/sync-media-pub/blob/master/manifest-exts.md#1-sync-mediachapter1json
Daniel Weck: it’s based on the concept of the EPUB 3 MO, i.e. a tree that more or less maps to the HTML tree
… and define mappings between document fragments and audio chunks
… fragments are defined with fragment identifiers
… there is no industry standard to reference character ranges except CFI
… in this proposal we mention fragment IDs and CFIs
… it’s designed to align with the OWP
… not using clipBegin/clipEnd in an XML syntax, but using the media fragment syntax
… the clock syntax is seconds.milliseconds
… the way we represent pairs of text/audio is to use JSON properties “text” and “audio”
… see the full example at the end of https://github.com/w3c/sync-media-pub/blob/master/manifest-exts.md
Ivan Herman: having played with SMIL, what surprises me is that we have to somehow reinvent the wheel, or at least redesign it
… it’s a Web thing, it’s not particular to Web Publications
Daniel Weck: it’s a CG, it’s not tied to Web Pub
Benjamin Young: +1 to targeting the Web for these things
Ivan Herman: yes, but you formulate it with the Web Publications use case
… we have to be careful, we don’t want people to think we’re forking the Web
… we have to put it at the Web level
Marisa DeMeglio: I think long range, having something like that on the Web is great
… but our personpower is very limited and our use case is very concrete
Ivan Herman: I’m almost sure that if we go with that as part of the recommendation then we’ll create a tempest
… we can’t avoid that to come at some point
Hadrien Gardeur: https://w3c.github.io/wpub/experiments/audiobook/
Hadrien Gardeur: I pasted a link of a basic experiment of an audio book
… as a Web Pub, and your position is saved using local storage
… a demo of what we worked on, where WebPub is directly applicable
… creating something that works for this on the Web is easy
… if you’re interested, reach out
… it shows that audio books is a low hanging fruit
George Kerscher: using pure audio and not sync media?
Hadrien Gardeur: right
Benjamin Young: https://github.com/menismu/popcorn-js
Benjamin Young: I wanted to +1 what ivan was saying that synced media is a general Web problem
… so we have to phrase it carefully
… the link I pasted was one such synced media format, using essentially timed-text like expressions in HTML
… a fork is somewhat active
Daniel Weck: Another implementation example (there are several): http://wam.inrialpes.fr/timesheets/
Daniel Weck: :)
Benjamin Young: I would suggest engaging with these people
… there are possibly still people at Mozilla who care about this project
… this was the Web-focused vision of it a few years ago
Daniel Weck: Also: https://www.w3.org/TR/timesheets/
Marisa DeMeglio: yes, it would be interesting to find out why they’re not pursuing this
Avneesh Singh: yes, it’s not a Web Pub-specific thing, sure.
… but the main use case is for the publishing world
… we want to have audio overlays without having to change a Web Publication
glazou: I know TimeSheets, kaze did it
glazou: loooong ago
glazou: obsolete IIRC
Avneesh Singh: we don’t need to put it as a Web Pub thing, but we need it and don’t want to get lost by broadening the scope
Daniel Weck: Also: http://webtiming.github.io/timingobject/
Avneesh Singh: we should focus on handling this use case first
… to address the issue of having it as part of the Web Pub specification, we can work in the CG
Daniel Weck: going back to Benjamin’s mention of popcorn.js
… I’m familiar with this, also a couple other initiatives like timesheets
… more recently timingobject, which creates building blocks
… I completely align with Avneesh’s comment there
… the main use case is sync media text/audio books now
Hadrien Gardeur: you kinda a need to alter the text for this to work, because we’re missing a good text fragment identifier
… for audio you can use media fragments
… for text currently you need IDs, which is less than ideal
… this goes way way beyond than synced media: bookmarks, annotations
Leonard Rosenthol: not to mention, web annotation selectors
Hadrien Gardeur: I’m wondering if that’s not an example of sth that our group can work on and make a difference
Ivan Herman: Web Annotations has two different parts, one is the annotation itself (data structure), and one is the addressing part
… the addressing part goes beyond what the Web Annotation need, we had these discussions there too
… the other thing is that I don’t think I disagree with Avneesh
… it’s all a question of presentation
… if I just talk about “Web page” instead of “Web Publication” then there’s no binding to this WG specifically
… it may be more politically acceptable
… the technology can be the same, defined in the CG
… my preference would then be to publish that as an independent specification
Daniel Weck: I agree with you Ivan
… half of the proposal is glued to Web Pub, but the other half is generic and can be split
Tzviya Siegman: what’s the next step? how do we promote this?
janina: we were really concerned early-on about these dragons mentioned by Ivan
… we heard Marisa yesterday, and it needs to be very clear that the existing technologies don’t meet the requirements
… we would need an explainer
Marisa DeMeglio: yes, these documents are recent and we want to make it an explainer
… what we would need is admnistrative guidance on where this work should happen
Ivan Herman: for the time being, doing what you do is fine, in the CG
… when you have a draft specification, we can see where to publish that
George Kerscher: it wouldn’t be included in Web Publications?
Ivan Herman: it could be…
… the core of it for me is this “mini-SMIL” has to be specified separately
… and then how this integrates in the Web Pub manifest is a different thing
Tzviya Siegman: I will follow up with Marisa and Daniel in the PWG
… it will be on the agenda in a few weeks
Romain Deltour: — LUNCH —
Garth Conboy: scribenick George
3. EPUB4
Garth Conboy: We have EPUB 4 followed by Audio books..
… Garth has some notes for this topic. EPUB 4 as a concept has more relevance to many of the members of the WG.
… We talked about this this morning in the business session.
… Because of the status of EPUB 3.2 ,, we are unclear about what EPUB4 should do.
… Before the discussion at business, EPUB4 was going to be a packaging of the WebPub.
… The boundary is special in web publications. If you are making an EPUB4 from a web pub, it must be bounded. The packaged version would have the right stuff.
… The resources in EPUB 3.2 is a feature, but other people see it as redundant.
… If we are thinking of EPUB 4 as having bounds, the question of resource list is interesting.
… Perhaps EPUB4 is not bring a lot; however if we look at what is lacking, the aubio book space is it.
… What is being received for distribution is all over the planet and it is ugly.
… Most of what we did from whole cloth was not adopted.
… If we look at audio books, we have an opportunity to spec it without retooling, they could ingest that content.
… Question is will the publishers adopt what we spec out.
Garth Conboy: suggestion is to explore we keep a package of what we have like EPUB 3.2 for audio books.
Dave Cramer: What sort of things should be called EPUB versus something else.
… Technically EPUB is the content, metadata, packaging. If we replace the OCF with something else is it still EPUB, probably.
… There is a more fundamental issue. There is the web model and this EPUB model.
… In EPUB, you take this content and bundle it and hand it to somebody for distribution.
Tzviya Siegman: +1000 to dauwhe
Dave Cramer: https://resilientwebdesign.com
Benjamin Young: +1 to dauwhe
Dave Cramer: Our description of WP falls between the cracks. We have WP and do what they do on the web.
… I wonder what we are doing with this distinction. Are we just bundling up to send to somebody?
Garth Conboy: if we change the manifest to JSON and bundle it up, it is still EPUB.
… We can correctly use the term EPUB for stuff that is packaged upp and sent out. Where WP is a web hosted thing.
Dave Cramer: web packaging may change that model. It is a potential bridge between the WP and EPUB.
… We as publishers want to use more of those web
Hadrien Gardeur: We see people using EPUB and it is not really meant for them.
… With EPUB there is a lot of baggage. Things that do not make a whole lot of sense.
… We could start from scratch and design for what we need.
Liisa McCloy-Kelley:
… I struggle to understand how we can resolve problems. The audio book is very timeley. We can solve problems for this fast growing area that has no standards.
Avneesh Singh: If I look from outside, it will be difficult to explain EPUB 3.2 and EPUB 4.
Benjamin Young: +1 to Avneesh
Avneesh Singh: Perhaps put EPUB 4 in a freezer and wait for a better time.
Garth Conboy: +1 to Avneesh
Avneesh Singh: Some markets will find WP to be very important to their segment.
… We can make an informed decision in a few years.
Garth Conboy: ?
Avneesh Singh: Audio books we can move forward on. We should have an audio book community group.
George Kerscher:
… Great idea for the audio book CG.
Tzviya Siegman: We need to move forward with an audio book spec.
… Iceboxing of EPUB 4.
… WP we don’t exactly know what the path forward is.
Garth Conboy:
… WP could continue in this WG or in a CG. It should continue in some way.
Tzviya Siegman:
… WP as it is today I will not implement.
Luc Audrain: EPUB 3.2 on a rec track will be proposed to the PBG. What will happen to EPUB 3.2 would come from the BG.
Hadrien Gardeur: Tech specs can take a long time. Implementations can take longer.
… I don’t know that iceboxing is the right approach.
… From a tech spec perspective, this not does not make sence.
Garth Conboy: EPUB 4 moving forward creates market confusion.
Ivan Herman: Let’s continue with WPub and eventually EPUB 4 would make sense. WPub may be appealing in the future.
… Audio books could be published directly, which would be a good thing.
… WP gives us a framework to work on a number of things.
Garth Conboy: WP continuing, epub 3.2 rec track possibly, and audio books as a special case of WP.
Garth Conboy: a?
Garth Conboy: q
George Kerscher: audio spec: rec track?
Garth Conboy: possibly an ouput of this group
Leonard Rosenthol: core responsibility bring pub-specifica aspects to the web
… help the web evolve, things special to publications, we want into core web platform
Avneesh Singh: working group, innovations in WP, such as audio books, gather feedback, leads to specification work
… rec track for EPUB 3.2: no need to fix all details now, need open view on what will be tested / implemented, etc.
Romain Deltour: can someone clarify re-chartering?
Ivan Herman: process-wise, replace current charter with new+removed work items etc.
… first this community needs consensus, then other members to support it (W3C -wide)
Dave Cramer: re-chartering opportunity to get objections to recharter again
… Combining IDPF into W3C: EPUB 3.2 not “exactly” a web thing
… make heritage of EPUB “web things”
Tzvia: ISO relevant too?
Karen Myers: conversations / outreach to broader community
Brady Duga: differences between adopters / stakeholders for EPUB vs. WP (audiobook market, etc.)
… audiobook closer to EPUB than WP
Ivan Herman: EPUB3.2 requires technical changes to support audio books?
Ivan Herman: EPUB 3.2 needs interoperability testing, etc.
Brady Duga: focus should be on packaged audio book
… starting from unpackageed (WP) to packageed (EPUB) causes production problems
… (challenges)
… business need: packaged audio book, no need to address online / web audio book
… (Brady’s view)
Garth Conboy: leverage work in WP audio books (not Web Packaging packaged yet / zipped), let’s not open can of worms of modifying EPUB to support this
George Kerscher: trademark audiopub book would be good (communication / marketing) to outreach to industry people who have this specific need
George Kerscher: incentivize other market segments to join the broader group
George Kerscher: EPUB3 Media Overlays already supports audio, but too complicated
Dave Cramer: open question about evolutionary potential of EPUB 3.x
… audio file MP3 could be spine item
… (e.g.)
Dave Cramer: EPUB 3.2 currently draft community group report, so can be changed
… if compeling reason that can be done without current framework, then let’s do it
Hadrien Gardeur: there is definitely a need for audiobook in WP
… stream audio books (not just public domain)
… commercial audio book market, no need for DRM, just need manifest to handle that part of the user experience
Brady Duga: I do not disagree
… There are two uses: WP and packaged, two different things / market needs
Avneesh Singh: charter scope broad, do we need to broaden?
… WP Charter would need removing EPUB4 removal, that’s all (need to re-charter?)
Ivan Herman: we could create separate working group for 3.2
… convince W3C management that charter interpretation is correct
Luc Audrain: +1
Ivan Herman: not a clean way of doing it
… keep one Working Group
Avneesh Singh: EPUB3.2 and WP different objectives, target audiences
… if we decide rec track for EPU3.2 for existing publishers
… WP different charter possible
Lisa: audio podcasts, not just “audio books”
… another community to reach out ot
Avneesh Singh: +1 Luc
Luc Audrain: I don’t see how for EPUB 3.2 this would make sense
Daniel Weck: (rec track)
Garth Conboy: it would be a recharter so that the group can work formally on WP and EPUB3.x
Ivan Herman: for example Web Architecture works on several rec tracks in parallel / work items
… maybe separate WP and EPUB concalls / task forces
… it is up to us to see how we proceed
Garth Conboy: a?
Avneesh Singh: agree. WP = more involvement of web community, dissociated from the EPUB work
… EPUB 3.2 to meet the needs of publishing world
Garth Conboy: this group is umbrella vs. some other organisation is a detail
Laurent Le Meur: EPUB 3.2 to rec track is not interesting for publishing business, and it is a can of worms
… audio books need package format, so why would zip be less interesting than Web Packaging format
… (short term)
Garth Conboy: Web Packaging is a long way off, right now we can use zip
Tzviya Siegman: if we split WP EPUB groups we will isolate brains
… lots of us work in ebooks,
… we need to talk to other groups (web people)
… e.g. benefits of Service Workers, etc.
… lots of technical overlap
… they will do work with us, but we need to present a EPUB evolution message
… stop thinking as business model driving factor for technology
Romain Deltour: +1
Luc Audrain: concern - how to sell this to the publishing industry
… EPUB 3.2 to rec track at the condition that this does not modify significantly the technical underpinnings (heavy investment cost)
Dave Cramer: fix reading system bugs in ecosystem
… talking to the rest of the web: TAG, etc. people interested want to help, but we need advice on high arch level, need to figure out how to communicate our needs
… solutions likely to come from them
… create an environment where they will want to help us
… system design / integration, EPUB so far from the web, need their help to get closer
… communicate what we want without implying we have a solution (can come across off-putting)
George Kerscher: can we engage them about rechartering?
Dave Cramer: interesting to have converstation about that
Ivan Herman: now convinced better keep group as single task force
… publishing community still seen by rest of web community as strange, so better preserve existing weight of working group to influence
… let’s avoid breaking down / splitting the group
… if there was EPU3.2 -specific group there would be increased isolation
Laurent Le Meur: my point was Business Group to study EPUB 3.2 on rec track
… not to advocate splitting
Ivan Herman: some want ISO
… transition from Community Group to ISO is horrible. Better through W3C rec track
… more effective
Liisa McCloy-Kelley: minimal viable product, for as long as we’ve had EPUB, this is baseline level of support (fonts, pagination, etc.) functional scope figured-out by reading system implementators
Ivan Herman: take EPUB features, demonstrate in reading systems
… show it is not doable on the web
Avneesh Singh: single group more difficult to manage, but better for integration
… EPUB3.2 rec track too early to make decisions
… kind of research project (ISO, etc.)
… too premature to make decisions now about rec track
Tzviya Siegman: IG had many task forces / “working groups”
… CSS has many documents at the same time
… idea of sticking to separating working group seems very expensive
… (management costs, time, overhead)
Luc Audrain: decision by end of november feasible?
4. Next F2F and other meetings’ schedule proposal
Ivan Herman: need to timetable next year’s meetings
… Laurent EDRLab Digital Publishing Summit
… and workshop on archival of web content
… PWG F2F meeting tentatively begin of May 6-7 Mon-Tues most probably Cambridge MA
… think about it :)
… week of June 3rd
… in Paris for workshop + symposium
… Cambridge MA maybe Google, Hachette, MIT
… Paris maybe BNF national library
… Cambridge MA is for FTF (not workshop + symposium)
… Boston marathon is week after
Daniel Weck: —- BREAK —-
5. Audiobooks
Wendy Reid: Last session … woo hoo!
… update on AUdioBooks (AB) task force.
… Current state: AB TF formed in May
… Focus has been on WP (not EPUB 4, to be named later)
Garth Conboy: — Working sample of audio-only AB
Wendy Reid: on Web and Android
… A couple of issues
… Hurdle to resolutions: lack of use cases; have worked recently on these
Daniel Weck: https://docs.google.com/document/d/1JZfSLE_cO1g3jgpHMx8Q99rQBwGyhOFwNWvfXAwW5Fg/
Wendy Reid: Audio WP has a player than move forward/back, can play, provide metadata, change playback speed, retain position
… online, offline, streamable.
… AB WP open-able useable in non-WP-aware browser.
Daniel Weck: https://w3c.github.io/wpub/#audiobook
Ivan Herman: on non WP browsers, the term is fuzzy
… JS usage these days makes this hard to address
Juan Corona: .. WPs with JS bridges problems on browsers.. but what category is this?
Hadrien Gardeur: yes, but you can listen to tracks and skip around
… but we need more than that, we need JS to do more
Wendy Reid: we have some posted issues, for use cases, here they are:
Juan Corona: Hadrien’s demo too
Hadrien Gardeur: https://w3c.github.io/wpub/experiments/audiobook/
Wendy Reid: https://github.com/w3c/wpub/issues/307
Wendy Reid: schema.org issues with duration, you need the total pub for the whole pub
Wendy Reid: https://github.com/w3c/wpub/issues/308
Wendy Reid: individual elements too, to let the user know the progression at the chapter
Juan Corona: Second issue: bitrate, format
Wendy Reid: thanks to the discussions yesterday, hopefully we can come to a resolution
… so let’s close some issues and continue
5.1. Duration of an audiobook #307
Juan Corona: (chatter among the group) technical details
Daniel Weck: Media Fragment URI uses NPT Normal Play Time https://www.w3.org/TR/media-frags/#naming-time
Ivan Herman: let’s stick to what we have now, because we ran out of time for the draft
Benjamin Young: latest Normal Play Time (NPT) RFC https://tools.ietf.org/html/rfc7826#section-4.4.2
Daniel Weck: … ISO 8601 not suitable, hoping duration from Schema.org will be ammended https://en.wikipedia.org/wiki/ISO_8601#Durations (affects totalTime too from Schema)
Ivan Herman: introducing our own element now is not viable
Wendy Reid: let’s close this issue
George Kerscher: not sure if we need the duration for each chapter
… there’s the use case but people just want to know the book duration
… everything else could be calculated along, from the files
Wendy Reid: but there’s a track, we do get them as individual files, sometimes. but we get them as tracks instead most of the time
… it’s to get the best UX
… I want to know I’m 5 mins into a 25 min chapter for ex.
George Kerscher: so you get them as tracks.. you process them.. but what do you need?
Wendy Reid: it’s more complicated than that though
Juan Corona: it’s not 1:1
Brady Duga: my question is like george’s, wendy is this reflecting what’s current now?
… duration in the metadata is one form for us, but it’s not accurate. It’s concerning to have it like this
Wendy Reid: let’s talk about it, how does a pub look like when it’s packaged
Daniel Weck: bigbluehat RFC 2326 is referenced normatively by Media Fragment URI, do you suggest we explicitly reference RFC 7826 instead?
Wendy Reid: duration is on the chopping block I know, but we need it for the UX
Brady Duga: ToC, chapters mapped to the range of the chapters/tracks… you need to know how to calculate it.
Benjamin Young: DanielWeck: RFC 2326 states it’s obsoleted by RFC 7826, so I’d link to the newer one
George Kerscher: audio file sequence, you have a pointer to the beginning or the offset.
Brady Duga: if you have the info before hand you can calc it. but in the web you don’t know this info until you download it all
Romain Deltour: its a metadata issue, not a playback issue
Brady Duga: george you are right you can calc it, but it’s hard when you are streaming it
Wendy Reid: we need to know where in the chapter you are, with knowing the duration and offsets mixed with aggregation of the chapters or not
Lloyd Rasmussen: do we need to have an API and the player to coordinate this
Brady Duga: would the publishers give us this detailed ToC information?
Wendy Reid: the spec won’t eliminate this missing data, or bad data, but we can validate it and provide feedback
Daniel Weck: there are different ways to organize the files, theres a case theres a chapter with a title,
… or more sophisticated with a ToC, and files contain more than one chapter
… have the information on the manifest, along with the link
… any objections to this?
Wendy Reid: close issue with duration, use npt, until we get a new direction
Daniel Weck: media fragments has a different reference
Benjamin Young: there is a new one
Leonard Rosenthol: is it compatible with media fragments?
Benjamin Young: depends on what you are using, if npt> link,, to reference the new rfc coordinate with that group.
Wendy Reid: media fragments for now then
Resolution #1: Close issue #307 (duration) after discussion with schema, we will use NPT (Media Fragments) until they confirm or inform us otherwise research needs to be done
Benjamin Young: general question, not on the TF sorry.
… it’s not essential for pkg, its for navigation
… why is it in the manifest?
… it looks like annotation
Juan Corona: .. could it be on its own
Benjamin Young: someone could share this info like an annotation
… I feel like this content is not essential to the manifest, it could be tripped on
Wendy Reid: this is in the proper context though, this is out leading solution for now
Benjamin Young: there’s too much going on on many spaces
… a track in HTML5 is something else
… HTML wants a playlist system, it’s missing
… this is a good use case, to converge with.
Daniel Weck: this is the reading order
… page-list, audio streams
Wendy Reid: any objections to close this?
Daniel Weck: who edited the draft?
Ivan Herman: this is not for me, tbd in a PR
5.2. Alternate formats or bitrate for an audiobook 308
Wendy Reid: https://github.com/w3c/wpub/issues/308
Wendy Reid: issue 308 bitrate, alt formats
Wendy Reid: no normative statement about which format MUST be used
… but we need a way to indicate what format is referenced (for streaming etc.) for reading systems / player implementations
… make decisions about what to download / cache
… use case for different qualities
… content creator to communicate bitrate, at the very least
Brady Duga: should this group handle this issue
… what about formats that have multiple bitrates etc
Ivan Herman: agree, should not discuss here.
Laurent Le Meur: alternative renditions of same content, DASH better because bundles multiple alternatives
… user agent to choose alt based on metadata expressed in manifest
Garth Conboy: what is the real-world use case
… in trade world
Wendy Reid: Audible does it, but not at distribution stage
Marisa DeMeglio: what’s special about audio books is that HTML mechanisms cannot be used in manifest
Hadrien Gardeur: only works for formats on audio and video (HTML)
… we have to address this issue if we want a solution that works everywhere
Brady Duga: if we allow this, we then need a fallback mechanism
… this is complicated
… create multiple version of the book instead of bundling / declaring the alternatives inside a single publication
George Kerscher: what about ingestion?
Wendy Reid: ingestion, we would be checking the files
Daniel Weck: Benjamin’: HTML audio video do not include bitrate, selection based on format
Wendy Reid: EPUB vs. WP audience, but I realized audiobook work predates all of this, venturing into solving problems that are solved at different level on web
Hadrien Gardeur: content negociation not likely to happen for audio files
… content management backend not smart enough
… some other group will have to deal with this, if this group / we don’t do it
Benjamin Young: https://www.w3.org/TR/html/semantics-embedded-content.html#dom-htmlsourceelement-media
Benjamin Young: not asking about content negociation, just the HTML media source spec
Romain Deltour: responsive image community group
… user agent to negociate, HTML picture element source media
… new HTML element + extended source element to allow for content negociation
… we could reuse constructs semantics , but would need to define processing model
Benjamin Young: this is the crux of the issue (processing model HTML vs. Web Pub manifest
Ivan Herman: -> https://www.w3.org/TR/html/semantics-embedded-content.html#the-picture-element Picture element
Wendy Reid: 308 need more thinking
5.3. Mixed media in audiobook linear reading order #322
Wendy Reid: https://github.com/w3c/wpub/issues/322
Wendy Reid: issue 322 mixed media
… audiobook readers don’t support supplemental content (sc)
Laurent Le Meur: WP can help supporting sc
Wendy Reid: the topic is important for the WG as a whole
… publisher will probably mix content in WP even if this is not stated good practice.
… an ad could be insterd in the middle of an audiobook
Hadrien Gardeur: my concern is that we won’t be able to easily identifiy such supplemental content
… the UA will discover it when it tries to fetch it.
Brady Duga: maybe not too bad. Some annoying cases will arise. And there is no processing model defined.
… A use case may be show a picture of a character + play the corresponding audio
… but without processing model, we don’t know what to do.
Juan: why don’t specify a duration with an image?
George Kerscher: we could put the image in the ToC.
… but it’s going in the sync media model
Ivan Herman: I don’t understand. Each item in the reading order has a media-type.
… the media-type may be wrong but specification wise, this is a fact.
Hadrien Gardeur: therefore for an audiobook it”s a requirement to add a media-type. I’m ok.
Marisa DeMeglio: this sounds like a case for sync-media. Pure audiobook and sync-media should have proper overlap so that audiobooks can get such bells and whistles.
Garth Conboy: the generic case is a business book with tables, and a pdf.
George Kerscher: we could do it with a ToC.
Wendy Reid: it conrresponds to a use case in which the ToC has images.
Ivan Herman: in fact in the current spec there is no default, I was wrong. Maybe we should say that a resource should be an html (by default), if no media type is given. In an audiobook each item would therefore require a media-type.
Hadrien Gardeur: the problem is only when there is mixed content. It would be reasonable that if a WP is typed “audiobook”, it should contain only audio files.
Proposed resolution: if a string appears in readingOrder/resources/links, the canonical manifest should include a Publication link for that URL signaled as text/html (Ivan Herman)
Wendy Reid: or if all resources are audio files the UA sees it as an audiobook and treat it differently than a generic ebook.
Avneesh Singh: Audio book may be a profile, which allows to say that it is recommended to have audio content in default reading order
Proposed resolution: if a string, or a PublicaitonLink without encoding format, appears in readingOrder/resources/links, the canonical manifest should include a Publication link for that URL signaled as text/html (Ivan Herman)
Brady Duga: you say that by default, even for an “audiobook”, the content is html. This would make a stupid format.
Ivan Herman: how to be sure the content is an audio file only from its media type?
Daniel Weck: we’re deviating from the web, where the media-type triggers behaviour in the UA
Ivan Herman: previous proposal resjected ;-)
George Kerscher: does epubcheck goes through an confirms the media type of content?
Ivan Herman: if there is no explicit media type there is no checking.
Daniel Weck: the same remarks we made about mixing content applies to comics also.
Benjamin Young: this discussion is not great for web browsers
Benjamin Young: to DanielWeck’s point:
{"@context": "http://schema.org/", "@id": "http://w3.org/", "@type": "AudioBook"}
is correct (but rather insane) JSON-LD
Daniel Weck: where to I get the information of book vs audiobook? the audiobook type in the manifest just states the schema.org schema will validate some metadata.
Laurent Le Meur: the meaning of “audiobook” should be stronger than that and indicate also that it contains audio
Daniel Weck: for sync-media we use “audiobook” with html content.
… I’ll open an issue.
Hadrien Gardeur: hint should not be trusted. You should prepare for fallback when something goes wrong.
5.4. relationship to minimal viable wpub
Ivan Herman: general question: we discussed about a minimum viable WP. Today we described a minimal viable audiobook. does it mean that we have a different minimal ebook vs audiobook reader?
Wendy Reid: the minimum features are the same.
Ivan Herman: the MVbook there is no notion that I know where I am.
Hadrien Gardeur: that was more about non WP-aware UA.
Garth Conboy: we have 2 or 3 must fo WP UA, seems they apply on audiobooks.
Daniel Weck: filed new issue: https://github.com/w3c/wpub/issues/351
5.5. next steps
Wendy Reid: next steps = draft sections that we do have, we have a working example.
… I want to talk to the industry, talk to the audiobook publishers.
Hadrien Gardeur: publishers will rather create the packaged version of audiobooks.
Wendy Reid: which package system, which metadata, these are the questions.
… they might have requirements that we didn’t see until now.
… after 5 months, we’ve gone far and it’s cool.
Ivan Herman: who will do the fist editing? related to duration etc.
Hadrien Gardeur: if the package format is for audiobooks, we may choose a specific extension and medi-type.
Garth Conboy: I think we should not use the token “epub4” at least for the moment.
Leonard Rosenthol: why not .pwp?
Ivan Herman: then we would have different flavor for comics etc. Is it good?
George Kerscher: if we go to audiobook people without talking about packaging, they won’t have a clue.
6. Resolutions
- Resolution #1: Close issue #307 (duration) after discussion with schema, we will use NPT (Media Fragments) until they confirm or inform us otherwise research needs to be done