Publishing Working Group F2F Meeting—Day 2 — Minutes

Date: 2019-05-07

See also the Agenda and the IRC Log


Present: Deborah Kaplan, Benjamin Young, Dave Cramer, Ivan Herman, Luc Audrain, Wendy Reid, Matt Garrish, Tzviya Siegman, Tim Cole, Ralph Swick, George Kerscher, Laurent Le Meur, Romain Deltour, Rachel Comerford, Avneesh Singh, Jun Gamou, David Stroup, Charles LaPierre, Franco Alvarado, Marisa DeMeglio, garth geoffjukes, Nellie McKesson, Toshiaki Koike, Joshua Pyle, Daniel Weck


Guests: Wendy Seltzer, Leslie Hulse, Jeremy Morse, Andrea Martucci, Jeff Jaffee

Chair: Garth Conboy, Tzviya Siegman, Wendy Reid

Scribe(s): Benjamin Young, Brady Duga, Garth Conboy, George Kerscher, Franco Alvarado, Ralph Swick


Tzviya Siegman: we’re picking up from where we were yesterday
… mostly focusing on building a Testing Task Force
… bigbluehat and Nellie mentioned contributing, but we need a Chief Cat Herder to help make this a reality
… I’m busy this summer
… I’m going to just start volunteering people

Tim Cole: I can also contribute, but not in a place to lead

Tzviya Siegman: do we have a project manager?

Rachel Comerford: sure

1. publication issues

1.1. CR Publications

Tzviya Siegman: k. we have what we need for that

Tzviya Siegman:

Tzviya Siegman: if you look at that link, you’ll find the timeline for our publications
… in section 2 these are our major milestones
… we’re supposed to have CRs in the 3rd quarter of this year which is rapidly approaching
… ivan can you explain more about CR?

Ivan Herman: we have to have a kind of feature freeze at some point
… before CR
… once we are in CR, we are not supposed to change technical parts of the documents or add or remove features
… unless we can also prove that something’s wrong
… so we should agree on a date after which we don’t make such major changes
… editorial changes after that will be fine
… but technical changes won’t be made unless something is uncovered by testing
… we do need a realistic timeline set for testing
… and counting back from that, we can determine a timeline for feature freezing
… so we should have CR published right after TPAC
… that would leave us way enough time to do testing
… we have to calculate to pass a recommendation we need to have voting time through the AC…which is 4-6 weeks
… so we have to calculate including that
… Septemberish/Octoberish is too early given time needed for testing
… I think that’s about it
… and we have 2 documents that go to rec: WPUB and audiobooks
… they cross reference each other, so they should go to CR together
… matt, do you have any thoughts?

Matt Garrish: feature-wise, we probably have a significant number of items at this point
… the concern in my mind is implementation details–as noted in the TAG review
… hammering those details out and getting people who want to implement these things
… we need commitments from implementers

Ivan Herman: there is a “feature at risk” option
… for things we’re unsure that this can be implemented or not
… we can formally mark things that way which are of concern
… if for any reasons, we need to make a significant technical change on something not marked as “at risk”, we have to essentially start CR over again

Dave Cramer: I think mattg already mentioned this, but we need to find commitments to implement
… and I’m curious to find out what things count as implementations
… if I write a JS-based reader, does that count?
… maybe if EDR Labs writes a JS-based reader, maybe there’s does count?

Tzviya Siegman: I’ll volunteer to meet with the group to define what counts as an implementation
… and talk to Ralph about that
… we also have a long list of open issues
… and those should be closed

George Kerscher: without content a web publication would be meaningless
… so a specification is meaningless without commitments from content providers
… how do we track that?
… who’s making efforts to bring publications to market
… what’s the story there? I’m not clear on that

Ivan Herman: strictly from a process point of view…well my perspective on that process…
… the goal of the CR phase is not to demonstrate usage at large or to serve testing implementations
… but to prove that implementation can be done
… this is partially an answer to dauwhe
… if there is a JS implementation which works somehow as an extension to the browser that is fine
… it proves that the document can be implemented
… some of these things happened in other groups and became part of communities’ tooks
… some don’t. I did some of this for the OWL spec
… but the goal wasn’t to make something for the community, just to prove the document’s content can be implemented
… marisa said yesterday that she almost had an implementation
… if it’s the only implementation, then its a problem, but it’s already a big step toward what is needed
… we do have some example books
… and we need to have examples for each of the terms we use
… to make sure they’re useful
… it is not a goal of the CR period to build up these examples, it’s only a welcome outcome
… we can put up goals that go beyond what CR requires
… but we have to be mindful that we are going beyond that

Tzviya Siegman: goal here is to get volunteers

Dave Cramer: I volunteer to do samples/examples
… I can say Hachette is not committing to publish in this format
… it’s way to soon
… I get what ivan is saying, but I feel that if that’s all we accomplish, that’s sad

Tzviya Siegman: we have a rough timeline for WPUB
… but for audiobooks, we have not published a FPWD

Wendy Reid: FPWD is scheduled for 2 weeks from not

Tzviya Siegman: there is additional feedback coming for audiobooks
… and hopefully we can get everything to CR on a similar timeline

Garth Conboy: if I was betting on anything right now, I’d expect audiobooks would be the most useful first
… can testing audiobooks test WPUB since audiobooks is based on WPUB?

Ivan Herman: each document feature should have at least than 2 implementations
… so audiobooks will use some of the WPUB features, some of those features will be tested by audiobooks
… but some features are ignored by audiobooks, and those features will not be

Tzviya Siegman: so we have 10 minutes to finalize packaging/distribution document and talking about what other modules we might focus on for things based on WP

1.2. packaging specification

Tzviya Siegman: what steps do we need to take to finalize lightweight packaging

Laurent Le Meur: I think we just need one or two sessions
… during a call to finalize the open issues
… from what I heard yesterday, what was the bigger issue–the origin of the package
… that issue is disappearing if we’re only distributing it–and not using it on the Web
… if that issue is closed, in a month it could be finalized

Tzviya Siegman: early June?

Laurent Le Meur: the end of June

2. Other modules

Tzviya Siegman: the next question is what additional modules we’d like to work on
… there are many suggestions for things we might work on
… but no volunteers to make those a reality
… nor clear businesses driving the “we need this to happen in the market”
… now’s your chance to express what you’re going to build
… archiving would be great

Jeremy Morse: so I have an action item to possibly be included in the WP standard
… I have some stuff sketched out last night

Tzviya Siegman: I’d like to help out, but I’m going to volunteer dkaplan3 because she loves archiving and I like volunteering her for things

Deborah Kaplan: I’ll be happy to help with archiving jgmorse

Dave Cramer: we need to be concerned about timing. If we’re kicking off new exploration for things that will be based on these CRs, then there’s a very narrow window here…

Ivan Herman: +1 to what dauwhe just said
… any other module that we look at will not be on a recommendation track
… it may later become a rec-track spec, but not right now

Tzviya Siegman: bigbluehat makes a good point that the Publishing CG is a great place for these things to be explored
… and we can begin ideas here in the WG and then figure out if they go to the CG

Rachel Comerford: I’m not sure what we’re looking for here exactly…maybe I’m not the only one?

Tzviya Siegman: when we come to audiobooks, we have something specific–where as WPUB is very general
… there are other areas where these more specific formats will drive needs for the general WPUB
… possibly using Linked Data for scholarly publishing, but don’t necessarily get used across all of publishing

Ralph Swick: this is making me wonder if declaring feature freeze for WPUB is premature
… perhaps audiobooks can help harden what needs to be in WPUB
… I understand why things are timed the way they are
… but if audiobooks is the only module, then it’s perhaps not showing the promise of wpub sufficiently

Laurent Le Meur: if we add more modules, then we may have to add more constraints–such as audiobooks only using audio resources
… will other profiles also come with new limitations? for instance, will scholarly publishing restrict this further
… why isn’t Wiley using WPUB now for scholarly

Tzviya Siegman: we’re using just HTML so far because it’s been sufficient
… we could use WPUB, but we don’t yet have a need

Ivan Herman: this question, and also rob’s comments
… if we look at the web publication spec…there are 2 major aspects to it
… there is all the metadata structure…the fact that we use JSON-LD,, some basic building blocks for ToC, etc.
… all these fundamental structures are things audiobooks will need anyway
… any other profile will rely on the same kind of things
… there are many things that go beyond wpub like previews, etc.
… and there is a separate section in wpub that describes what a profile is and what they can do
… that’s mostly been about terminology/vocabulary
… we could do this for scholarly, but it also has strict rules about citation
… that may be part of the profile for that
… profiles can also make further restrictions on the format
… like audiobooks can only contain audio files
… so, back to what Ralph said, audiobooks would exercise some of the wpub features
… but would also need its own set of tests
… we need to now look at that profile section and figure out what it requires and what it opens up for changes to the base spec
… doing one for scholarly would follow a similar pattern to audiobooks
… it may introduce changes to reading orders
… comics world would have something similar if we made a profile for that
… I don’t know Rachel if it becomes clearer

Rachel Comerford: so these feel a bit like use cases for different business verticals within web pub
… so under that umbrella we have a format and use cases per vertical?
… do they live independently of CR? or do they become part of the WPUB CR?

Ivan Herman: they live independently

Rachel Comerford: shouldn’t the first generation of these use cases then come from the business group?

Romain Deltour: +1 to Rachel, we need business cases rather than theoretical specs

Ivan Herman: I’ll leave that to those chairs

George Kerscher: the DAISY consortium has been doing structured audiobooks for years
… so our request is that these new audiobooks should be able to maintain structure—toc, pagelists, etc.
… they seem possible, but there aren’t requirements on ToC, nor requirements on reading apps for using those things
… are these then part of future profile things?
… we don’t want to see audiobooks take a step backward to “just a playlist”
… or this will be rejected by our community

Romain Deltour: +1

Matt Garrish: the web publication is 3 parts now
… audiobooks builds largely on the first part and the third part
… there is a line that gets you from WPUB to audiobooks, but I don’t think it’s critical
… so it’s possible we could just define the manifest format, and then focus on audiobooks
… since it’s not required that audiobooks be a web publication

Wendy Reid: serious discussion happening
… we also have the UCR document we did for the audiobooks spec
… and that UCR document has all those use cases George mentioned
… but any implementer that wants to implement audiobooks must reference the UCR

Tzviya Siegman: much of the feedback we’ve gotten is that pulling stuff into the UCRs has confused how implementers will happen
… so, we may be pulling some of this back into the main documents
… we’re discussing that in the next session

Ralph Swick: differing things into UCRs won’t get you interop–which I think was George’s concern
… I’m concerned about feature freeze and brand
… maybe think about feature freeze on precisely what is needed for audiobooks
… and that WPUB 1.0 is that foundational framework…and the focus is on it having just what audiobooks need

David Stroup: I’m still catching up on this WG, but one of the things I’m wrestling with is actually what this all is
… what is the web publication supposed to do
… is it strictly the metadata?
… and then its up to implementers to do all the things George is asking for
… but we don’t actually say that they have to
… just here is that information and if you’re going to do that, this is where you find it
… so for educational publishing, audio and textual content is very important
… and moving between them is important
… and I keep hearing that audiobooks are more restrictive as a profile
… but is it actually any different than an educational profile?
… there seems to be lots of gray area

Avneesh Singh: Ralph said much of what I want to say.
… completion of WPUB is important, but I think its good to focus on audiobook
… I see WPUB as a process
… maybe if we’d taken more time to discuss the business cases
… we may not have heard that publishers aren’t planning to use it
… perhaps we will recharter along future profiles
… but for now, let’s make WPUB a framework, and have a solid specification of audiobooks
… and move on to do all the work following that

Deborah Kaplan: Avneesh did some of this, but the things folks have been saying for the last several minutes boil down to…
… we are finishing WP
… we have a schedule for it
… we also hope to eventually do modules
… some might happen now, some might happen later
… is that what I’m hearing?

Ivan Herman: yes

Marisa DeMeglio: what is the difference between a module and a profile

Wendy Reid: there isn’t one

George Kerscher: do these go through CR?

Ivan Herman: no. the audiobook is the only one going through CR because it’s happening now
… maybe in the future we do them
… but as of right now, any other profile is just informative

Benjamin Young: The profile and module term is most confusing because it isn’t really profiling
… unless we distill it down to just a manifest structure
… because we throw a lot of that out in audiobooks
… audiobooks are a different animal that also uses the manifest
… it’s not really a profile, it is a fork
… we could just ship audiobooks and not use wpub
… don’t need discovery, entry page, etc

Wendy Reid: You might want to read the audiobook spec
… we kept a lot of that stuff, and it is not optional
… we think there will be audiobooks on the web some day

Garth Conboy: I wanted to get back to dkaplan3’s summary
… it may be at odds with what Avneesh and George said before
… if we took Avneesh and George’s approach, then we would ship more of these and keep WPUB alive somehow
… but that’s different than what dkaplan3 said
… I’m not sure I have a strong opinion about either route, but maybe now’s the time we pick one

Wendy Reid: laudrain is here, and he’s chair of the business group

Luc Audrain: we have a chance with audiobooks
… something that will be implemented and will be used
… this the chance we have to create something that may be real for the business
… then it will build confidence in this foundational spec
… and then begin to consider future profiles
… so I think we should focus primarily on audiobooks

Ivan Herman: the audiobooks on the Web conversation frustrates me
… the whole procedure by which you get to the manifest which could be in the HTML file
… or at least linked from that HTML file
… you essentially unpack the zip file and start that process
… the reading order content is restricted to audio files
… the ToC may be included and is in HTML
… but it may be streamed or whatever
… it’s blatantly incorrect to say audiobooks are not related to WPUB

Deborah Kaplan: what is our actual plan for how we ship these?
… I have opinions on that
… but I’m more concerned that we’re not driving this to a resolution
… I thought WPUB was just shipping as the foundational underlying structure
… which is easier to bootstrap into systems
… but I’m now concerned that’s in question wrt to its relationship with audiobooks

George Kerscher: tell me where our accessibility checker fits in this
… I’m going to talk about scholarly publishing and audiobooks
… it would be great to know that audiobooks can come with some amount of a11y features expressed in the metadata
… same would be true for scholarly publishing. does this conform to WCAG AA, etc.
… I want to be able to have a11y conformance and discover those features
… is that a module?
… or is this an a11y checker we write that applies to WPUB and everything based on it?
… we have this in the publishing industry now, and I don’t want this to take a step backwards

Tzviya Siegman: I believe that right now the answer is going to depend on whether a publication is one file or multiple files
… if you have a single file, then we have things like Axe which work on single HTML files and we’re good
… if we’re working on multiple files, then we’ll have to adapt it to use the WP ToC document and then all the other multiple documents
… we want to change as little in the modules as possible
… so the ToC is the same in WPUB and in audiobooks
… the only difference you would have to make right now
… is that it goes across multiple object modules
… fundamental question, though, is whether we’re keeping WPUB open or closing it

Tim Cole: the audiobook seems to have support to go to CR on the current schedule
… WPUB doesn’t seem ready in that timeline
… education, for instance, would need changes to the WPUB foundation
… so WPUB seems to need more work to support these future modules
… perhaps we say, here’s audiobooks, it uses a foundation which we hope to use in other future profiles
… and as we do future modules, those modules shouldn’t break audiobooks
… so there’s a commitment to the WPUB core framework
… but it leaves it open for growth and change
… but if we take WPUB to CR, we won’t have that option
… when we explore future profiles

Geoff Jukes: this is a bit off topic, but kind of to address George
… at Blackstone, we have audiobooks, and a separate DAISY product
… if we can maintain that separation, then I feel adoption will be more likely
… and we could up-sell folks to the DAISY one
… and this notion of being modular becomes very appealing
… we can then sell the media-sync layer
… which can be plugged into the product without having to sell or distribute a separate thing
… this kind of thing could be really appealing to the business

Tzviya Siegman: so a lot of publications have an “extensibility mechanism”
… maybe that’s all we need?
… at least if its defined well enough?

Ivan Herman: that’s Part 3
… but in this side conversation I’ve been having with mattg
… it may be that we have to chop the document into separate docs–which I’ve been trying to avoid
… we have a general metadata structure
… we have a set of metadata items
… there are disagreements among the editors whether accessing the manifest is fundamental or not
… to make the whole transition process possible, we may have to make some of these separate
… to make the core possible to go to rec-track
… and see how things evolve later
… I don’t think we can come up with a real plan right now
… but that may be a possible avenue

George Kerscher: I think modules and profiles are actually different
… if you have your core, then ever profile must conform to that
… but then there are modules that can be added to create a profile
… if we create a module that does a11y reporting, then that could be added to the scholarly profile
… or to the audiobook profile
… video is not a thing in every publication, but if there were a video module, then perhaps that could go into educational publishing, etc.

Matt Garrish: is it important to delay WPUB because we’re concerned about future modules?
… other things may offer other features
… is it an overabundance of caution that causes this concern about taking WPUB to rec?
… or could we move WPUB and audiobooks forward together as originally planned

Wendy Reid: modules is not a real thing
… we’re not currently structured that way, and unlikely to be
… so do we proceed to CR on both? or just audiobooks?
… and leave WP open and add stuff to it as we go forward
… that is still the decision we need to make

Deborah Kaplan: so I agree completely that’s the decision to make
… but my opinion is that there are no blockers to this choice
… there is a revision mechanism in the W3C
… we can use that

Avneesh Singh: I don’t think we still understand the situation. we’ve been discussing this for about 2 years
… what is the business requirement?
… the charter says we have to publish WPUB
… so we should go through that and plan for potential changes in the future
… maybe just one call more to find things to soften
… which could allow audiobooks to move forward to CR

Ralph Swick: so I probably contributed to the bi-annual existential crisis talk this morning
… are we going to be able to test everything in WPUB?
… the art here, will be saying “WPUB is the foundation for all these things”
… the first thing based on this framework is audiobooks
… so testing audiobooks, should test parts of WPUB
… one technique that could be used is marking only certain things as normative
… and leaving undecided things informative
… which could be made normative later
… so, normative bits are the things audiobooks require

Tzviya Siegman: +1 to Ralph

Ralph Swick: and anything we’re not wanting to test right now become informative

Wendy Reid: +1 ralph

Charles LaPierre: +1 to Ralph, which also means we could go to CR for both WPUB & AudioPub

Ivan Herman: the only comment I had was to what wendyreid said, is that we cannot publish the audiobook by itself
… because it must have normative references to wpub
… so we’d at the very least have to copy/paste much of wpub into audiobooks

Laurent Le Meur: one question first…when?
… starting septemberish?
… what are the foundational conditions for that?
… one key issue is open since a year about alternative renditions
… we can’t freeze functionalities until we have at least closed that
… if we speak about web publications at large, things for ebooks, not audiobooks
… one thing never discussed was fixed layout
… many publishers will never use web publications if they don’t do fixed layout
… so, do we hope to leave that open to being discussed?
… and other things that have never been discussed?

Dave Cramer: CSS actually has position: absolute
… which does make fixed layout possible
… we should not be afraid of CR

David Stroup: +1 to move forward with WP and revise as needed

Deborah Kaplan: dauwhe +a gazillion

Dave Cramer: the most valuable thing we can get right now is finding out how these work out in reality

Tzviya Siegman: dauwhe++

Marisa DeMeglio: so I like the idea of a web pub as a framework
… one is a book
… so by using one or more modules, you end up with a profile
… that seems very clean and also extensible

Garth Conboy: there was something very strange that happened way back

Proposed resolution: Take WPub to CR, with normative content for anything that’s required by Audiobooks. Take Audiobooks to CR. (Garth Conboy)

Garth Conboy: Ralph made a proposal which got some +1’s
… this proposal (reworded by me) seems like what Ralph said, and I think we can +1 that for real
… Ralph is nodding

Tzviya Siegman: +1

Charles LaPierre: +1

Romain Deltour: 0

Garth Conboy: +1

Wendy Reid: +1

Ivan Herman: +1

Benjamin Young: +1

Deborah Kaplan: +1

Tim Cole: +1

David Stroup: +1

Jeremy Morse: +1

Garth Conboy: (portions of WPub that are not required by audiobooks are informative in first CR version)

Nellie McKesson: +1

Luc Audrain: +1

Ivan Herman: sometime around TPAC…we’ll have to have a separate existential conversation
… but we’re good at these
… the question is, what do we do after next June?

Romain Deltour: if WPUB 1.0 is only used normatively for audiobooks
… we say we shouldn’t be afraid of 2.0
… but if 1.0 is only used for audiobooks, why not just ship a single audiobooks spec
… what is a web publication if it’s only used for audiobooks?

Dave Cramer: +1 to romain

Romain Deltour: so why not just postpone it to 2.0?

Ralph Swick: I think it’s just a statement of timing, and that the door is staying open to future uses that the group has explicitly deferred

Ivan Herman: besides the scary editorial issues, I fear other profiles (like comics) won’t look at it
… so we’d go back to what we have today later?

Romain Deltour: yes

Deborah Kaplan: ivan +1

Ivan Herman: my instinct is that this would cause more harm than good

Resolution #1: Take WPub to CR, with normative content for anything that’s required by Audiobooks. Take Audiobooks to CR.

Garth Conboy: resolution means portions of WPub that are not required by audiobooks are informative in first CR version

3. finalizing the UCR document

Wendy Reid: Over to Franco…

Franco Alvarado: Reading over TAG review.

Franco Alvarado:

Franco Alvarado: Req. 17: A Web Publication should be able to express the access control and write protections of the publication.
… Req #17: WP should be able to express rights & access info.
… What if lose permission? Warnings?
… Req. 5: All constituent resources, and their contents, should be identified by either a URL or a unique handle that can be mapped to a URL.

Tzviya Siegman: TAG review of audiobooks:

Franco Alvarado: Concern: how does this interact with Web Security model — require origin.

Ivan Herman: -> requirement 5:

Franco Alvarado: Need to clear up TAG-review related issues.
… What items need to come through from WP?

Dave Cramer: Req #17 is DRM not far in disguise.
… Out of scope for WG, but perhaps in scope for BG.
… Remove #17?

Luc Audrain: Agree with dave.
… Perhaps BG or out-of-scope.
… Library lending systems.
… Agree, expunge.

Franco Alvarado: UC 69: A library may loan the publication for two weeks or a university may make a textbook available for its students for the course of the year. A Web Publication should provide a means to inform user agents about the availability period to enable the UA to control access accordingly.
… Yes, DRM in disguise. Okay to remove.

Luc Audrain: Link to privacy policy. That might be fine, but DRM is not WP stuff.

Tzviya Siegman: UCR:

Ralph Swick: Should just document this as “out-of-scope”, rather than just expunge.

George Kerscher: Some purchasing mechanism would control access to the WP (even though that detail is out of scope).

Franco Alvarado: Req. 5: All constituent resources, and their contents, should be identified by either a URL or a unique handle that can be mapped to a URL.

Tzviya Siegman:

Dave Cramer: WP resources mapped to URL — how does that fit with Web security & origin?
… Addressed on the Web today, but get more complicated with multiple layers (e.g., cloud reader).
… problem needs more explanation.

Franco Alvarado: UC 18: Judit uses an annotation tool to comment on a publication authored by Pablo. She puts an annotation against a sentence in a particular paragraph, anchoring that annotation to the sentence using a reliable way of identifying it. That identification should not be invalidated by a subsequent change of the document by Pablo (unless he, e.g., removes that sentence).

Luc Audrain: UC #18 — annotations — solved by Web Annotations? But not adopted.
… hopefully possible in the future.

Rachel Comerford: Is UC #17 a response to the CFI issues with EPUB for Edu?

Tzviya Siegman: maybe
… a UC can still exist if a solution exists.

Franco Alvarado: should said solution be included in the UC?

Ivan Herman: “a unique handle that can be mapped to a URL” — those UC’s don’t seem to map to that.
… Perhaps just say “URL” rather than unique handle.

George Kerscher: I agree with Ivan and the DOI resolves to a URL, which preserves the security.

Wendy Reid: should implementations that we’re looking to achieve be in the UCR doc?

Tzviya Siegman: Spec needs to have implementation clarity.
… How will be make implementation details clear?
… Need to make things clear for UAs.
… What goes where (UCR, Spec) for implementation details?

Dave Cramer: Confusion around UC doc — perhaps a gap analysis? What’s normal Web stuff?

Ivan Herman: +1 to dave

Dave Cramer: many UC are satisfied by Web stuff.

Tim Cole: UCR references the Spec — easy to get lost. Some examples, perhaps?

Ralph Swick: +1 to citing where a use case may already have been addressed by other web technology

Tim Cole: Challenging thing to do.

Ivan Herman: I love Dave’s (gap analysis) idea.
… avoid perception we’re reinventing the Web.

Franco Alvarado: Address this explicitly in the UCR document?

Ivan Herman: A table might be good.

Wendy Reid: AI for gap analysis? Franco and Josh likely best?

Action #1: produce a gap analysis of the ucr items vs. existing web concepts (Franco Alvarado)

George Kerscher: If an example exists in, say, EPUB, should we just point to that example?

Tzviya Siegman: perhaps. Real world is good, but not the same as the Web.
… EPUB is closed ecosystem, needs to be adapted to Web.
… What does an implementor do? We’ve pointed to UCR. Butk folks seem to want one document.

Charles LaPierre: Nice to have an example to be guided by even if its not specifically based on the web.

Luc Audrain: Good question from George.
… EPUB is about Publications. Gap analysis should also include EPUB 3.
… BG should review UCR document. Will be a BG AI.
… documents what exists and is practically useful.
… What’s solved in EPUB 3 is relevant.

Tzviya Siegman: how do we document implementation guidelines.
… RS Conformance from the EPUB days!

Ralph Swick: Heard reference to normative guidance to implementors to be found in UCR document — that doesn’t seem wise.
… UC #24 — listening to audiobook, would like to see chapter info — that’s really advice to implementers. Link back to metadata needed to address UC
… (in Spec).
… should where in Web suite where what’s required to implement UC’s can be found.

Ivan Herman:

Ivan Herman: Similarity to annotation model document — describes a vocabulary, not how a UA behaves.
… Like us.
… For each metadata item, shows small UC that uses it.
… Examples are good.

Marisa DeMeglio: Another great informative resource would be code snippets.
… Copy and paste is (generally) good.

Deborah Kaplan: Some UC’s will create implementation advice for AT’s.

Dave Cramer: What kind of code samples? Browser folks in C++, or polyfill folks in JS?

Tzviya Siegman: UC’s to identify Specs they are served by… and code samples are good too.

Geroge: If we’re providing implementation guidance — we have problems across bounds (multiple reading order items).
… Should these two approach to the search problem be included?
… is that obvious or should such be included in implementation guidance.

Charles LaPierre: Lazy Indexing or Background Indexing is other options as well.

Franco Alvarado: We’ve been encountering problems with varied audiences. Browser would or otherwise?
… Code samples are another possible mismatch with multiple audiences.

Tzviya Siegman: E10L crisis #2.

Ralph Swick: UCR’s are generally advice to WG or spec developers… not implementors.
… What have we failed to provide? Don’t need to fill each gap.
… “We’ve design section 4.3.x to meet this UC.”
… Better as advice to “us.”

Matt Garrish: Data-driven specification — in EPUB we merged UC’s into Spec — perhaps successful, perhaps not.
… Deal with what implementors want, and avoid to many layers.

Joshua Pyle: Code snippets are good. UCR is (should be?) a business documents. Code shouldn’t be there; they should be in spec, if anywhere.

Luc Audrain: +1 to josh

Deborah Kaplan: No religion about where advice to implementors lands, but needs to be somewhere.
… E.g., handling of cover art — implementors need a clear instruction flow of “if you want a cover in this view, get it from here, get the alt-text from there”
… Needs to exist somewhere clearly — deal breaker not to do so.

Dave Cramer: Guidance for implementors… but who are the implementors?
… Likely not native browser support.
… Perhaps a WP library to import in folk’s own projects.
… using just JS and existing APIs. Really a RS script. Kinda like Mathjax for WP.

Ivan Herman: Setting up unattainable goals for this WG?
… There should be an implementation guidline document.
… What we expect an implementtion should pick up this metadata and bounded content and present it as a uniform whole.
… Regardless of native browser or Polyfill (as implementation).
… We expect XX, YY features to existing based upon available data.

Joshua Pyle: It’s sad if we think browsers won’t natively implement this.
… Browsers want our content.
… If the content comes, Browsers will play.
… Josh remains hopeful.

Charles LaPierre: +1 I agree :)

Luc Audrain: +1 to Josh, again

Tzviya Siegman: need to come up with UCR action items… and how to proceed with implementation guidlines.

Deborah Kaplan: Not sure who are we writing implementation requirements for — we should say things like the information about a cover or its alt-text should come from here (specifically).

Benjamin Young: If we build it, they will come — no they won’t — hasn’t proven true with the semantic web.
… build as web apps first.

Tzviya Siegman: dkaplan3 had good exmaples — covers.
… Best practices, not UCR.
… Are we beating a dead horse? All else goes into Implementation Guidlines?

Franco Alvarado: A few more things need to get done. UA display info and WP authoring info. That’s the current state.

Ivan Herman: A higher level implementation vision should be a different document (or even Wiki page). Maybe not long.

Franco Alvarado: should that be informed by testing task force?

Tzviya Siegman: different animals.

Joshua Pyle: I don’t think we should merge branch that puts technical implementation details into UCR. It should instead be start of
… implementation document.

Tzviya Siegman: Looks like we’re wrapping up UCR document for publication soon.
… Implementation best practices effort to start — Matt & Tzviya.

George Kerscher: Not really best practices, really implementation vision.
… How our bounded content should work.

David Stroup: Is the UCR document a living document?
… as impacted by profiles?

Franco Alvarado: UCR tries to be generic as possible, not specific to profiles — but may well become more specific over time (with more profiles).

Tzviya Siegman: The UCR is a note, so can be repubished as frequently as desired.
… “final for now”
… Gap analysis still TBD and may have some impact.

Ivan Herman: Automatic publishing not set up for Notes, so a bit of a pain. But, can still get done.

Tzviya Siegman: UCR document — the whole thing is informative.

4. Addressibility

Tim Cole: Addressibility
… First, sections in editors draft

Tim Cole:

Tim Cole:

Tim Cole:

Tim Cole: Address, identifier, andprimary entry page,
… We do not have the actions quite right, wo there is work to do.
… Reading from the definition.

Tim Cole:

Tim Cole: We need to discuss the use of identifer

Tim Cole:

Tim Cole: has property value
… issue about hte identifier as a schema identifier that we’re using here. example 49 uses key id. i dont think we’ve talked about putting that in context
… its not the same @id
… we can have id to current key or identifier to both canonical identifier and for other types potentially that dont use property value, like DOIs
… 2 more issues: talked yesterday that one of the current concerns about the light weight packaging format which is the base issue. another potential issue:
… should it be possible to add a doi to a lwp
… addressibility of components and spans and components within a wp, particularly multi part wp
… for example: audiobooks, previews, identifiying spans
… a lot of other useful specs out there ( css selectors, shadow dom, eg) . we may not need as much of the web anno as we thought
… basic question: what other specs to reference? what do we need to invent with regard to addressibility

Benjamin Young:

Tim Cole: perhaps it should be a tech working group note that could later come back
… this goes back to cfi and its use to epub community. is it useful? if so, it has to be addressed in wp
… is there a problem of consistency between PEP and canonical identifier?

Matt Garrish: definitely a problem there

Tim Cole: how do we resolve it?

Matt Garrish: its just removing two paras. I can do that real quickly

Tim Cole: what about the issue of multiple addresses for entry page?
… canonical id has to be unique. pep also has to be unique. if that is true, they are the same thing. but that means pep can never be moved

Benjamin Young: pep is practically going to be called index.html
… if it is default, it already gets two identifiers.
… as slash and as index

Dave Cramer: slash = pub address, index = pep
… is there precedence to this?

Benjamin Young: yes, scope in service workers

Dave Cramer: how does this fit in with DOI?

Tim Cole: idea of canonical id was at one point palce for dois. we could drop canonical and just talk about identifiers
… the address is closest to identifier. the pep separate as index.html. doi, use the identifier property
… identifier could be lots of things. urls have to be urls. but theres more nuance to that

Ivan Herman: I may be too worried about this slash story. That means imposing certain structures to how author would put out their thing. Which is quite a tall order
… if i take the doi example. i just checked it with one of my doi. that resolves to a url with slash. but thats not a requirement that anyone imposes.
… therefore the doi could resolve to something we would not consider canonical identifier. i would be cautious
… the world may not do that

Jeremy Morse: +1

Ivan Herman: i understand slash is usually what is used.
… i think we build in a dependency from the outside world which is way too heavy.

Dave Cramer: this require ment may be a help to us. the naive web server you have a lot of stuff in a folder. this prohibits two peps in the same folder
… which i think eliminates the possibility of a lot of bad things. and especialy if these things have service workers to enable offline. could have different scopes…
… it seems to be an idea that could promote hygiene

Benjamin Young: practically it is how web devs do it now. not just apache. always some default response from something edning in slash
… index.html is not the req here. just make sure when you get the pep
… its not a hefty requirement or abnormal. solves for there being an index and an index-2.html and it being a second pub
… theres no technical way to say these dont conflict and theyre not entangled into each other
… by scoping it to essentially a director ending in slash, we avoid a host of problems. already how static site gens do it. we’re still avoiding issuing statements like ….
… must be called this or that. still not dictacting urls

Tim Cole: not sure i agree with ivan. i think that we prob should not decide today. id like to see this discussed in the issue and PR. there are gonna be use cases we havent tlkaed baout yet
… ex: i have something that sreves as page one of article and pep of wp, but is also one of several articles thatm ake up a journal
… ben should propose this and talk about it later

Jeremy Morse: two points - i dont think this is anticipating platforms where several wps might be served nested under a site and served by an api
… more generally – yoo don’t want this to only be adopted by only new or existing wps. requiring anything more than add the manifest …
… is just another further barrier to adoption of the spec. which seems like a problem. if it already looks like a wp, then slap on the manifest and its a wp

Dave Cramer: ivan how would you disambiguate between a url of a whole and the url of the pep itself ?

Ivan Herman: i dont have an answer. but this answer does not satisify me.
… i would be more happy even if its also a set up problem if we say something on the http response instead of the format of the url
… because the url is just something that is too narrow in this sense. http response might include info about this is the wp or something like that
… i think we are buying ourselves a problem if we impose a format of uris
… maybe in a few years youll get urls that have a different syntax
… like DIS or whatever, and suddenly we are stuck because we are stuck and we imposed a certain url
… we drive ourselves in a corner with such a very strict constraint

Benjamin Young: there are ways to avoid this concern. all of them have a path delimiter.
… more vaguely: end your path but not specifiying a specific file. goal is to disambiguate between the wpub and the pep page itself.
… .. we shouldnt fear those futures. if they become popular, that wp 2.0.

Dave Cramer: i think this is worth proposing and seeing what the reaction of the community is. i want to game this out
… we might learn something interesting.
… i think trying this and seeing what people think is reasonable.
… starting out strict, its easier to relax requirements

Franco Alvarado: .. than the opposite

Laurent Le Meur: when someone makes a package before something goes to the web …
… when we go to def of canonical id, there is no preferred version of the wp at this time
… there is a solution to avoid canonical id and wp says this is ok.
… but this means the process means this must add canonical id
… other solution is to use doi. but this is not good, because you need the url
… we have to solve that and recommend something to creators of packages and its not evident at all

Garth Conboy: sounds like benjamin is going to do this as a proposal.

Action #2: file issue about publication addresses being required to end in / (Benjamin Young)

Benjamin Young: this doc

Ralph Swick: -> Web Annotation Extenstions for Web Publications

Tim Cole: we have a first public working draft of web anno extensions for wps. ratther than push that to recommendaiton, can we hand that off and develop more slowly?

Ivan Herman: the clean way: this will not go through a rec track right now. i think the proper way that groups do, that htey publish this as a note
… and hten from that point on , if you are at iannotate, talk to the people there, if there is an interest. there is work going on with media sync where these things can be tried.
… basis for experimentiation
… once we put it in a note, we make it clear that at the moment we are not working on this as a potential recommendation right now.
… my proposal would be to publis hit as a note and for the time being consider as closed within this working gorup

Tim Cole: id like to clean up a couple of loose ends if we pub as a note

Benjamin Young: how many of you knew about this document?
… everyone should look at whats in here.

Proposed resolution: publish the Web Annotation Extension for Web Publication as a WG note (Ivan Herman)

Tim Cole: +1

Ivan Herman: +1

Benjamin Young: +1

Matt Garrish: +1

Romain Deltour: +1

Dave Cramer: +1

Resolution #2: publish the Web Annotation Extension for Web Publication as a WG note

Tim Cole: lw publication is a longer discussion.

5. issue bashing

5.1. Ensure text equivalent for accessibility

Wendy Reid:

Wendy Reid: accessibility in audiobooks

Avneesh Singh: audiobooks: good for many people but people with hearing loss and maybe some other disabilities, audiobooks doesnt fufill accessibility requirements
… for all streamining media there should be a text equivalent. this is soemthing which makes audiobooks inaccessibility. we had this discussion. the direction we are going in is …
… we should be having a provision of having some kind of pointing to text equivalent
… there is a link that goes to audio file. at the same time, a link or some other mechanism goes to an html page or sync narration file
… this is the proposal on the table
… we yet have to figure out details
… but this is the main concept on the table which can be resolved .

Garth Conboy: is this a requirement or something you need to do ?

Avneesh Singh: there was a concern that this will increase overhead. but actually this is the definition of horizontal review. but it sstates that there should be a WAY of achieving it
… they need this to make wcag aa, aaa

Marisa DeMeglio: what we are discussing on this issue is to use a porpoety in the manifest like alternative which could point to an html file, for example. or could point to sync media file
… this same poperty alternative we could use to create sync media content. it doesnt make our spec any bulkier and satisfies horizontal review

Deborah Kaplan: question: it looksl ike this proposal if im reading it right, does not allow for programmatic identification that this is an alternative taht can be parsed by AT. it seems like this is just a link to an alternative and then the onus is on some other human or aspect of publication to make it clear as to what the alternative is

Marisa DeMeglio: we havent gotten to the point of high level metadata like epub has. this property could point to a resource that has a mimetype identified. so you know if its pointing to a file. but it doesnt give an overall perspective

Benjamin Young:

Wendy Reid: we should have a flag that says that an audiobook has text assistance

Ivan Herman: two ways to look at this and i dont know which one we want. one is to say: this is a propoerty that refers to an alt to the audio. in which case this is a property that is defined for audiobooks only.
… which means that this should be defined in the audiobook spec. that is sort of straightforward
… question is there a more general concept here.
… do we want to get more general about alternative methods of displaying content? or do we just want to get it right for audio and worry about general later

George Kerscher: in epub accessibility spec, weve got these things like confroms to wcag aa, or a, etc. audiobooks have been used for accessibility purpose for a long time. it seems crazy to
… identify an audio book as not being accessible. what we did in that spec is identify a publication that is optimized for a particular group. aduiobooks fit into that category
… so it would conform to the audiobook spec in that metadata
… its reasonable if we do point to an html version of that book, that you may have to buy it. i think in the metadata in we should say that its not accessible to people that are deaf so access mode would be audio
… and in the accessibility summary we say this is an audiobook. we should say that its available as a wp or epub. on the practical side we do those things in the metadata
… and solve our rights problems
… audiobook pub doesnt always have text rights.

Marisa DeMeglio: great points. i think that having a property to indicate an alternative would go in a primary deinftion of the manifest

Wendy Reid: we do address the accessibility symmary in the spec

Benjamin Young: for linked resource, we already give rel alternate for free. we dont need an alternate propeorty
… related to this: if its per audio file then we’re doing each one of these texts and theres a need to provide more than one thing, we should accommodate that. if we can bring in the linked resource pattern,
… we can get a lot more things for free inthat space rather than a new property

Deborah Kaplan: going back to ivans original question
… . id just like to argue that making specific decisions about audiobooks where we treat accessibility particularly for audiobooks is probbaly a bad ideas. given that there can be accessibility issues …

Benjamin Young: +1 to not doing a11y stuff for just audiobooks

Deborah Kaplan: there are a multidute of media which will require alternative content
… its safer to say that we are not going to treat an accessible alt to audiobooks as a special case; we should have a generic solution

Matt Garrish: it reminds me of mutiple rendition of an epub. it becomes complicated quickly. when we’re gonna have to spell out what these alternatives are. its going to grow. they didnt get much traction in epub….

Avneesh Singh: agree with ivan about generic alternatives. i am not too much tied up whether it should be file level or top level. we just need a solution
… im not getting in to specifics.

Ivan Herman: maybe that is what benjamin have in mind but i did not understand it. if we do in general then we have the problem of somehow saying what is the alternative for.
… what can be done in the current sturcture and trying to change as little as possible…

Laurent Le Meur: Issue is a proposal for alternate resources in a link object

Ivan Herman: is to realize we do have already now the possibility of listing links which are reseooucres in the book but are not useed directly. if in an audiobook which is in a reading order
… i say alternative to audio, the html file must be in the list of links because it is a resource
… that means there is a linked resource object for that alt file
… then we need a vocab maybe. yes we can use rel attr for that which says this is what it does
… worry is do we have that vocab? how do we express it? i see mechanism in terms of metadata structure
… but there is a point where we need to fill it with specific terms

Laurent Le Meur: already a proposal for that. one year old now. iss 308.
… there is a resolution of that that has been made in readium
… link to what has been done in the readium context. simply an alt array of links.
… already used for audiobook alternate bit rates.
… it can be very flexible

Ivan Herman: this is not exactly the same i had in mind but we can discuss separately. this is different. this is alternate bit format. in original case this is a text alternate
… information should be there
… it is implicit because of media type
… is it enough thatw it is implicit?
… or is it somehting thta has to be expresssed explicitly? i dont know
… if media type is enough im happy

Laurent Le Meur: it is sufficient for RS devs

david: regarding media type, in my expreince its sufficient for it to be on audio. youll get alt bit rate, text alternate, etc. we do that quite often.
… it sounds like we need being able to express alternative both at file level and pub level

Deborah Kaplan: it seems like based on all this discussion, 44 gives us some of the info we want (uri, media type, but not the very useful a11y vocabulary features)
… in a generic way

Wendy Reid: audiobooks needs a mechanism to make it accessible to users. what we need to decide: how do we do that?
… main proposal is alternate on linked resource level
… two options: at resource level or pub level?
… but at pub level, looking at metadata to say whether this book is accessible with an alternative
… do we go ahead with linked resource alternative or do we go with a different function?

Ivan Herman: what is in the proposal just to make it clear is indeed ok as it is (see
… this is what can be done easily. requires one single additional property: alternate with value of url
… for the time being it seems that no extra info is needed because the media type is there
… thats fine

Benjamin Young: if we dont use audioObject, how do we justify that?
… already a lot of this in audioObject and associated media is close enough. if not sufficient we should make a claim that audioobject needs something else , if it isnt caption or transcript
… more general request that we need to make sure we are consulting before doing new stuff

Wendy Reid: it looks promising

Tzviya Siegman:

Ivan Herman: i dont see any problem whatsoever

Benjamin Young: we should take a look
… lots of stuff in that could be relevant. lets wring out

Wendy Reid: should we explore this as a solution?

Deborah Kaplan: the vocabulary is more complete for accessibility than Ivan’s proposed solution from #44, it’s true.

Ivan Herman: this comes back to my question: if we say audio object, it has a transcript. which means if we say in the audiobook spec, any resource listed should also be a typed out audioobject, then we get transcript and issue is solved, right?
… 44 is solved without having to do anything
… the question about whether we need a more generic version it is pushed back because we have resolved the audio specific issue

George Kerscher: are we requiring that to be in the audiobook, that there be a transcript

Wendy Reid: no. we are giving that option if they want to
… we dont have a mechanism for it at all, aside from making html files supplemental content. gives a publisher that option

Ralph Swick: [Wendy noted that schema:transcript only takes text values and audiobooks would likely want to be able to provide a URI]

Action #3: review “AudioObject” schema and explore what feedback is required for schema (Wendy Reid)

George Kerscher: an image without alt text, as long as we provide the ability to provide alt text, we’re good citizens and our spec provides the facilities to do things right even though we know the audio publishers dont have the rights to the text
… i do think that the whole metadata about audio publications needs to conform to the accessibilt conformance and discovery spec. otherweise we are doing a disservice

Wendy Reid: we may have also acceidently solved 308 as well.
… audio object has a thing for bitrate and encoding format, etc, that are in issue 308.
… i will review to make sure it fits all our needs and whether we need to give feedback to schema

Laurent Le Meur: an array of of media objects or html objects, this alternate array becomes a union of different json ld types which works. my only concern may be the author of the manifest who writes somethign by hand as dave says
… to add a type this type that on each of these things is extra work. the faciliity of only mimetype is lost because there are mor things to add. i dont say no but maybe it will be a bit hard

Wendy Reid: yes we’ll have to look into that as well but i think it brings us closer to resoving this issue

Marisa DeMeglio: i do appreciate that there are some things in schema that can solve our problems, not just #308 or #44. but we would still need to intro a propoerty to describe sync media to linked resource. so we might end up with both solutions whereas just having one term would be more elegant
… i dont think we have to solve it today. just want to get it in the minutes for my own future recollection

Ivan Herman:

Ivan Herman: there is a slightly more general thing because there is the mediaobject which is the superclass whcih ahs other subclasses like image object which is probably of interest to us for the same reasons. if wendys exploration is positive then there is a general statement
… there are a number of types that you may want to add.
… in the audiobooks we may have that as a requirement that you must use audioobject. i think we shouldn’t forget about that

Benjamin Young: linked reosurce: any plans to send to ?

Ivan Herman: we had a discussion started to talk about what features we need. is not known to be fast in responding. for the time being it is there. we would prefer to go the way but these are what are missing.

Benjamin Young: there is some way we could imagine it as a subtype of media object. but its good to know its been attempted

Tim Cole: we talked to to get bibliography. once they got the ideas right. we want these 25 things. they do have a pattern for this. theyre slow and they like it better when you bring stuff to them.

Ivan Herman: the chairs and I have already internally discussed that we should have a horizontal review. the problem is that the only name that i know could be helpful is rcihard wallace. i dont know who else could be helpful. i think this is something that in the near future we should send our documents to
… should have a good contact

6. Business group

Ivan Herman: slides ->

Luc Audrain: Update on the PBG, EPUB 3.2 has been approved
… EPUBCheck 4.2 has been released and a lot of texting has been done and bugs have been resoved. It has been translated to six languages.
… EPUBCheck is more precise and it has resolved issues for publishers.
… EPUBCheck is more accurate with EPUB 2 as well. This may help to move publishers to EPUB 3.
… ISO, a new Accessibility 1.0 is being moved to ISO. It has been rewritten using the ISO language.
… It is important for adoption, because governments need ISO specs. Also in Europe there is a need for ISO standard for accessibility
… Luc: EPUB 3.2 has overwheliming been approved by the CG and the PBG. Thank you Matt, Dave and everybody for the hard work.
… who wrote the accessibility act knows about this work in ISO and we want them to refer to this ISO standard.The EU commission
… Events where we can talk about our work BEA is one place, but hard to know how easy it is to get a stage at Book Expo America (BEA).
… June 25 and 26 there is a good opportunity to communicate.In Paris at the publishing summi
… In Japan Fukuoka Japan in September during TPAC there will be a publishing conference.
… Juc says Daihae has spoken to publishers in Japan and plan specif items for the 2 and half hour publishing meeting. This will help to move publishing in Japan.
… Action items are commucating about the events.
… This publishing victorys are an achievement. EPUB 3.2, EPUBCheck 4.2 the movement to ISO standards are all great achievements.
… We need testamonials from implementors in support of 3.2 and EPUBCheck.
… We want all members to push information out through all their channels.
… Finally, EPUBCheck fund raising has been good, but we are stuck at 2/3 of what we need. The next phase will start and finish in December, but we need more funds to complete the work we plan to phase will

Tzviya Siegman:

Luc Audrain: Avneesh is running in the AB election. We want to get the vote out for Avneesh. Get your AC rep and AC reps you have access to to vote for Avneesh.
… At TPAC there will be demonstrations.
… A new CG will internationalize and develop markets for a range of products.
… What will happen in Spring of 2020 in the BG.
… What will be the roadmap will be the topic.
… There is the issue that new releases of the print and it is the same ISBN, but just no edition change. The EPUB is also being updated as well. It is the same work and only the packaging is updated. Same goes for accessibility. We know how to make them accessible. We will not change the ISBN, and how do we let the users know. The users need to know that a new better file sis available.
… We are improving EPUB files based on the new EPUBCheck. Files correct years ago become invalid with EPUBCheck 4.2.
… The PBG should review should review the issues that this brings forward.
… if somebody came to his door and said a book he has has been updated.Dave: Publishers are interested internally. Dave would be annoyed and
… We get automatic updates in Apps and they don’t tell you enough? I wonder if the is a spec or process issue.

Tzviya Siegman: very few readers want the update.

George Kerscher: I have heard the Japanese Manga publishers are not interested in developing their product.

Luc Audrain: The Manga market is digital. It is a great success. We must be careful in this area.

Dave Cramer: working on final steps to publish EPUB 3.2
… what should the EPUB CG do next?
… considered Best Practices but lost an editor
… may get back to that
… some cleanup to do for EPUB for Education
… parts fo the IDPF spec have shifted to IMSGlobal
… we need to clarify what’s absorbed into EPUB 3.2 and what’s abandoned
… discuss with BG the future path for EPUB

Tzviya Siegman: +1000 dauwhe

Dave Cramer: how does this work relate to WPub
… high-level discussions on where all this might go to help guide the technical discussions

Luc Audrain: see “business cases” on my slide
… the BG should provide input
… I’m hoping that the release of the audiobook spec will build industry attention

Dave Cramer: dinner conversation last night on how business cases influence technical decisions
… I’m concerned about the business case for Web Publications

Romain Deltour: +1

Dave Cramer: that’s what’s behind my earlier statement about Hachette US and Web Publications being far outside Hachette’s business model
… possibly W3C might not be the venue for this

Jeff Jaffee: the question of right venue for business disussions is an important one
… we should ask the BG participants and the Steering Group to discuss that within their companies
… if we have a new spec that may be important in some business we need to know who they are

Tzviya Siegman: I hope the BG can help answer those questions
… I’ve heard mention of EPUB 3.3 and wasn’t aware that was in consideration
… this WG is chartered to work on “EPUB4”; maybe we need to talk about what that is

Dave Cramer: who is the parent of EPUB4?
… EPUB 3.2 or WPub?
… I do think there’s room for EPUB’s evolution
… do they open the door to new types of publications?
… so we’d need to do cost/benefit analysis
… things like cleaning up old cruft
… and it would be cool…but doesn’t it help people reading books?
… but I’d like to find out
… some of the questions around how to monetize things on the Web
… for trade publishers, that’s a very scary topic
… and audiobook folks seem scared of standards
… but if we take standards conversations too far into business, then we run into a different problem

Luc Audrain: I didnt’ mention EPUB4 in my slides
… I think the BG should work on business cases, not specific standards
… if there is a business case for trade books they will happen
… I hope audiobooks will anchor the success for WPub
… education and scholarly publishing are also business cases

Benjamin Young: what is the business case for WPub?
… scholarly is already on the web
… I’ve heard more reasons for the success of EPUB
… the thing we’re all excited about in this WG is audiobooks
… which look like EPUB-in-JSON
… we have two routes to the same [set of] features
… I don’t know what WPub brings that Publishers are saying they want
… maybe they will in the future but they’re not now

Rachel Comerford: I agree with Benjamin
… I can’t explain what WPub will bring for Macmillan

Romain Deltour: I want to remind us of the Publishing CG

Tzviya Siegman: +1 to CG

Romain Deltour: the web is built from incubation
… some things have taken a long time to incubate

Tzviya Siegman:

Romain Deltour: things don’t happen by magic; the CG succeeds only if people bring resources

Luc Audrain: as BG co-chair I don’t want to forget the business
… there is an EPUB business today
… our energy should go into Web Publications, not EPUB3
… there may be a need for an EPUB 3.2.1 with small tweaks
… Web Publication and its profiles are the future
… it may take a long time
… we have identified priorities
… the Publishing CG can incubate technical questions

Romain Deltour: BG should care about staffing + long-term strategy, putting the resources where they are useful

Luc Audrain: I’m not afraid of the future; I think it will take time
… many publishers are still publishing with EPUB2
… we need to tell them to use EPUB3
… about Web Publication, I agree with Dave that content protection is an issue
… but there are many EPUBs that are not, or no longer, protected
… these need to become Web Publications

Laurent Le Meur: I disagree that audiobooks are EPUB-in-JSON
… there are two parts to our work
… EPUB3 is not good for audio content so we’re doing something else
… but to expose audiobooks on the web we need to expose it in pieces due to their size
… I was myself surprised to see that Readium has several projects, including a ReadiumWeb – a piece of javascript to do Web Publications
… this shows that reading on the Web does have traction
… I regret that scholarly publishing won’t use Web Publications soon
… but there will be books with analytics inside

Tzviya Siegman: we’ve had this discussion many times
… Web Publications 1.0 won’t be perfect but we know it will be used by some
… let’s stop bemoaning that 1.0 won’t be perfect

Dave Cramer: the glory of the Web was enabling many more people to have a voice
… HTML was so simple – it lowered the barrier to people
… EPUB enabled self-publishing
… lowering the barrier whether people were monetizing their work or not
… reading books on the Web is still crappy; the experience is still sub-optimal

Deborah Kaplan: +1 dauwhe. I know multiple free, not for profit publishing platforms who want enhanced publishing functionality.

Dave Cramer: there’s a lot we can do; not sure exactly what it is
… we should keep trying even if our use cases dont’ work for $1B publishers

Garth Conboy: I view that we’re doing exactly the right thing
… we got a long way into WPub as a standalone thing then realized that we needed the audiobook profile to help evolve it
… we’ve not done anything dumb
… whether EPUB 3.4 makes sense is a separate discussion

7. closing thoughts

Tzviya Siegman: closing thoughts …
… we’ve recorded lots of action items in these two days
… there’s a lot of work to complete those
… the WG will meet on Monday
… we’re on track to publish
… we have a group of people to work on testing to prove that we do have implementations that work in the real world
… we’ll reasses in a few weeks

Jeff Jaffee: this team seems pretty confident that there is a worthwhile use case to proceed with
… that’s good
… I also hear angst that it’s not good enough
… I’d like to suggest that the WG assign to the BG and/or SC to do all they can to help create the business case and help popularize it
… the BG has focused on epubcheck which was necessary

Ivan Herman: +1 to jeff

Jeff Jaffee: this would be a great thing to do next

Tzviya Siegman: Romain reminded us of the Publishing CG
… we should make a strong effort to make sure the business cases and the technical cases are aligned

Rachel Comerford: +1 Jeff

Ivan Herman: it’s still far away but I think it’s OK to think early about the rechartering period
… we’ll need to seriously consider it in 6 months
… be more serious about getting the BG and WG to work closely together
… to help us avoid e10l crises every other meeting
… in January 2020 we’ll have to start thinking seriously about a new charter or an extension of our current charter
… we should make [the cooperation with the BG] much better than it has been

Avneesh Singh: with EPUB 3.2 just out there’s a gestation period for adoption
… making decisions before adoption would be premature
… we might waht to take 6 months [in 2020] to observe adoption
… the BG can concentrate on what are the highest priority use cases
… see what is adopted
… take 6-10 months to decide next priorities

Tzviya Siegman: the overwhelming advice from the AB/Publishing meetup in New York was “these things take time”
… “work with the developers and implementors”
… we might have a really good spec but are taking the wrong approach
… we can’t expect instant delivery
… That’s All Folks!

8. Resolutions

9. Action Items