Meeting minutes
Publishing process
Wendy Reid: Welcome!
Wendy Reid: 2 days of meetings, we do have a decent amount on the agenda
… We also have some guests joining us
… Hopefully we can be productive
Brady Duga: this is a WG meeting, isn't it?
… so do we have same rules for IP and non-disclosoure?
Wendy Reid: yes
Wendy Reid: Boring stuff first
… Talk about publishing process, there are some questions from the last meeting
… [link to slide deck coming]
Matt Garrish: Going over the W3C process for how we make changes
… There are 4 classes of changes
… Class 1 - no content change (maybe CSS, etc)
… easy, just publish when you want, no W3C approval needed
… Class 2 modifies the text in some way, but with no normative change
… Things like typos, punctuation, maybe remove a note
… Also does not require approval from w3c
… Class 3 updates/clarifies/changes normative statements
… for example epub:type clarification of elements it applies to
… mgarrish has lost audio
… This requires approval
… Class 4 adds new features
… This is pub manifest and audiobooks only, not epub
<Matt Garrish> https://
Matt Garrish: There is a list of substantial and non-substantial changes
… we need to tag every change with these levels
… After it is tagged everything works automatically
… for class 3 and 4 changes, they need to be annotate (change log isn't enough)
… Need to make a diff with ins and del tags
<Matt Garrish> https://
Matt Garrish: See link for an example
… Change logs need to detail substantive changes
… Only current changes are expanded, the rest are collapsed
… Class 3 and 4 needs approval, but CR and PR, etc stages
… Basically do an AC review, AC votes, then we remove the diff markup and publish the new text
… WG decides when to publish
… Looking at every 6 months
… If we need to we can release faster
… Notes are unchanged, same process as before
… changes are just done
… new notes still require approval
Ivan Herman: There is a long chain of current and previous recs, this needs to be done for various legal reasons
… tooling is missing at the moment so it is a chore
Tzviya Siegman: The sound is awful
Wendy Reid: Do you have PRs open for us to review
Matt Garrish: There are a couple
<Matt Garrish> w3c/
Open PRs
w3c/epub-specs#2572
Matt Garrish: Add "none" as an option for a11y
… Better than just missing statement, since it has a clearer intent
… This is a class 2 change
… combines informative pieces into new informative section
Leonard Rosenthol: Can someone clarify the change from "may" to "can"? Does it matter?
Matt Garrish: This is a lowercase "may", so it is not considered a normative statement in our spec
<Murata Makoto> I have to leave now, since I have to attend the plenary meeting of ISO/IEC JTC1/SC34. I will be back at 15:00 UTC.
Gregorio Pellegrino: The issue was raised by the federation of European publishers
… We can say several things about the a11y of the publication
… It can be 1. Checked and it is ok, 2, checked and it is broken, 3. did not check
… Adding this information allows us to better understand whether the file was checked or not
… This is important for statistics to report to lawmakers
… Also needed to align with onix
Wendy Reid: Any other comments?
<Wendy Reid> Proposed: Merge PR 2572
+1
<Ivan Herman> +1
<Matt Garrish> +1
<Wendy Reid> +1
+1
<Shinya Takami> +1
<Charles LaPierre> +1
<Masakazu Kitahara> +1
<Tzviya Siegman> +1
<Avneesh Singh> +1
<wolfgang> +1
<Ken_Jones> +1
RESOLUTION: Merge PR 2572
<Dave Cramer> +1
<Matt Garrish> w3c/
w3c/epub-specs#2571
Matt Garrish: EAA has 3 exemptions
… onix has 3 values for these exemptions
… We want to match onix
… New note with metadata field that can be used for EAA
… but also works for other jurisdictions
<Bill Kasdorf> Could you repeat the three values, please?
Gregorio Pellegrino: These are less technical and more legal
… The three values are good to cover these
… There was a request to add annotations to the values, you are free to do that but we are not adding anything specifically
… There could be a much larger quantity of these reasons
Wendy Reid: Librarians might need that, but we probably don'tt that much for epub
Matt Garrish: Doesn't seem that important to add now, we address as needed in the future
Gregorio Pellegrino: Just want to add that Christina has reviewed the note and has given her endorsement (she is very involved in EAA)
<Ken_Jones> @bill three values are The official names of the EAA exceptions are: Microenterprise, Disproportionate burden, Fundamental alteration.
<Wendy Reid> Proposed: Merge PR 2571
<Bill Kasdorf> Thanks
Ivan Herman: For the publication of this note, we have to wait for TPAC to settle
… And there is a week when I am on vacation
… We should set Oct 3 as pub date
Matt Garrish: Do we have to decide if we are using echidna?
Ivan Herman: We already decided to use echidna
<Matt Garrish> epub-a11y-exemption
<Ivan Herman> shortname: epub-a11y-exemption/
<Wendy Reid> Proposed: Merge PR 2571, publish a new note for the Exemption property on oct 3, short name will be "epub-a11y-exemption"
<Wendy Reid> +1
<Ivan Herman> +1
+1
<Tzviya Siegman> +1
+1
<Matt Garrish> +1
<Shinya Takami> +1
<Avneesh Singh> +1
<Ken_Jones> +1
<wolfgang> +1
<Toshiaki Koike> +1
<Masakazu Kitahara> +1
<Bill Kasdorf> +1
<Charles LaPierre> +1
RESOLUTION: Merge PR 2571, publish a new note for the Exemption property on oct 3, short name will be "epub-a11y-exemption"
<George Kerscher> +1
Wendy Reid: Do we want to do the third one?
w3c/epub-specs#2527
Matt Garrish: That is old IDPF stuff, and we discussed bumping something down
… Not sure if we can do that
Ivan Herman: It is still hanging in the air
Matt Garrish: First we need to resolve whether we can change old docs
Wendy Reid: Sounds like we need to talk to the system team?
Wendy Reid: Unless we have further publish process comments we can break early
… in 25 minutes we will have the next session
… we will start ON TIME for the next session ( we have an invited guest)
… reconvene at 3:45 Seville time
<Ivan Herman> guest+ sebastian
Anti-counterfeiting
The presentation slides are on line.
Sebastian: growing concern fraud retail platforms print and ebooks
… fake accounts may sell content they dont have the rights to.
… possible approaches were reviewed.
<Bill Kasdorf> Will these slides be available?
Sebastian: tried to dev. a structured approach. will share the results.
… not any legal evaluations, just the technical specifics.
… news reports from publishers in local markets, italy, germany and usa
… entire books republished with fake accounts, attach fake metadata, author info etc.
… there is paywalls to get additional content.
… repos fake accounts offer own content, but use the authors data, best sellers
… original content is modified from the author's version
… fake accounts, book promoted as a best seller but the book contains empty pages but had the same cover author etc.
… selling reprints was another issue.
… publishers had to identify multiple platforms, why should the publishers do this, not the retailers?
… we need to identify these fake books and be proactive to remove these fake items.
… if we can identify a similar cover image could be a red flak. we need to determine the original version. Make public and verifiable reference to aid in identification of the fake items.
… Human generated vs. AI generated is another concern.
… visible or invisible messages into the media file. Manifest embedding, certificates, into assets metadata C2PA initiative.
… verifiable identification, metadata, these compliment each other and if used in combination can be a benefit
… funded by the EU commission.
… support our work on verifiable credentials. this just recently started 1 month ago.
… verifiable content, metadata and rights to the work. 2nd reinforces this with public credentials. cryptographic identification.
… ICC originated in German Book market in 2017.
… ICC drafted , ISO publishing early next year.
… example of ICC, created directly from the digital media file
… generating an ICC can be done by anyone with access to the file and uses cryptography.
… different versions or file formats, or content with metadata, will be similar crypto hash but will be able to determine they are different but somewhat similar.
… changing the color of an image will still be able to find a match cryptographic hash,
… near duplicate versions / clusters can be determined even if the content is changed slightly. this will help to find these fake digital files.
… decenteralized registries can be used to look up these original rights holders.
… Publishers can contribute to a robust network to help detect fake media files, and submit their own ICC hash codes for authentic original covers, metadata etc to these decentralized repositories.
… we ask you to join the anti-counterfeit task force to help us with this effort
Dave Cramer: I bought a book from Norton, I realized it was counterfeit because of the quality. the one displayed was the cheeper version.
Sebastian: this comes up often. I don't know what to say. best solution would be content recognition, if the same cover is used by two accounts and choose the correct book publisher.
liisamk_: Amazon is trying to resolve this. they are finding things but this is spreading fast.
… building a trust chain from end to end.
… we need to programmatically fight this.
Tzviya Siegman: 1. blockchain is complicated. Most book publishers are slow moving, 2. generative AI is a big problem.
… detecting that is very difficult.
Brady Duga: We did think about blockchain to break out of walled gardens, was very complex, privacy issues we decided not to go that route.
… Dave and Liisa wish you were here to continue this. Sebastian, this seems very technical, hackers will figure out how to circumvent anything you do.
… this is a complicated as there may be multiple versions all being similar trying to determine fakes from the original, but hard with multi nation entities with multiple imprints who may also have the rights.
Wendy Reid: who owns the rights, easier from Penguin Random House, vs some random@gmail.com publisher.
… we need retailers to step up when two items come in with the same name red flags should be raised.
… we do have to think about the self publishers, we need it simple for them to use as well. CORESOURCE would be a part of the solution here.
… cover can be unique, images in the book can be stolen.
Leonard Rosenthol: other assets, 3D, video, mp3 etc. looking to figure out generative AI. multiple groups involved. to determine parts of text to the entire publication being AI generated.
Ivan Herman: we are the publication Maintenance WG, is there something that the EPUB standard has to adapt itself to changes/additions to work with any of these various solution. Or is there more we need to do technically. I am worried that EPUB file as 1 major binary file, its more complicated than that. how do you apply the whole coding env. with all the files inside the EPUB. What this WG should do vs. laywers need to do ect.
liisamk_: Shows an example.
… metadata is slightly changed. cover may be changed
… but this is not PRH selling this book.
… either we will tell Amazon this is a fake, or Amazon will find this and remove this title.
… would be easy for a small publishing house to consider this. the blockchain is used for immutable verifiable transactions.
<Dave Cramer> MD5 (9780774867740.epub) = 761ef29ed57c3d01101d660c9d46ef3d
Leonard Rosenthol: To Ivan Herman, Maybe, whatever solution we come up with or multiple solutions we could require changes to EPUB we leveraged certain file formats and may require changes to bring to those groups. yes there are changes are required but where that happens is another conversation.
Sebastian: these codes can be generated from the EPUB files, and we are working with creating these hash codes that make it easy to do and very fast. Huge amount of development make it to utilize the block chain without the need for crypto currencies / wallets / NFTs etc. so you wouldn't even know you are using the blockchain.
… verifiable credentials has a lot of use cases.
… we need to see what that technology can do for us.
Wendy Reid: What does it mean for this WG? either ICC / C2PA codes etc. how do we express them in EPUB. Is it even appropriate or is this outside of EPUB. Content creator and content reseller. WE need to consider that. if we want it in EPUB does it make sense. how would this work in publishing. Are retailers willing to do this. Are they going to implement this.
Tzviya Siegman: Who is hosting the blockchain, what is footprint of the blockchain.
… is there an expense to this.
<Leonard Rosenthol> ^^ Thanks @Wendy Reid
Sebastian: there is no need to embed ISCC codes in the EPUB. the interesting part is ICC code is external / no code/ id/hash into the EPUB. you just use an ICC generator. We discussed using ICC codes from publisher to retailers, to aid in deduplication, etc. we didn't use any blockchain it was just peer to peer.
… everything is proof of steak, not proof of work so no env. impact.
… provides a timestamp and your signature and you can sign your public key and your wallet address can be certified.
digital signature is a huge deal, and you can digitally sign your creative works and others can verify it is really you.
Brady Duga: I worry the issue here is source of proof of rights of publications are private contracts and as long as that is true, how can you make verifiable claims. author gave the rights to another entity who may also have the rights to sell that book. there are n number of private contracts
Tzviya Siegman: we must be missing something. a demo would be helpful.
<liisamk_> only since the dawn of ONIX...
Dave Cramer: expressing digital rights is complicated.
Sebastian: demo 1 min. contracts representing the rights holders, and they have the address / wallets.
… Shows an example. I generate the ICC code and can be tested.A plugin can be used to generate the ICC code and the fake image found on X, shows the original with the real credentials vs. a fake one
Ivan Herman: what is the next step for Publishing Maintenance WG?
EPUB/A
Tzviya Siegman: I now introduce Alicia Wise for this proposal on EPUB/A
alicia: I have a background in libraries and archive
… we collect publications from all around the world and our goal is to archive content for future
… I would like to speak about long-term preservation
… it's important to preserve publications, in history we've seen cases of burning books to "delete" history of people and cultures
… ISO maintains standard for preservation of different formats of digital publications
… I've been asked to join a group in ISO for thinking about a standard for archiving EPUBs
… these ISO standards are important internationally, because there're trusted by organization all around the world
… there are different kind of audits for publishing the spec
… a trusted archive needs clear specs, open, transparent
… we also have registries for knowing what content is stored where
… this is the context in which we're thinking about an EPUB archival format
… in ISO/IEC JTC 1/SC 34/JWG 7 we have a working group that is creating a subset of EPUB 3
… with features that are suitable for long-term digital preservation
… identifying metadata elements for catalogs
… we are a rational document, we have a candidate statement for EPUB/A
… the preparatory work of EPUB/A has been done on old version of EPUB
… so I think it's important to stay in touch do align with the latest version of EPUB
Shinya Takami: who is the target of the EPUB/A format? end-users?
alicia: the users are the trusted archives
Shinya Takami: so publishers should produce both "normal" EPUB and EPUB/A
alicia: yes, publishers normally deposit publications in trusted archives
… so they should produce an EPUB/A for deposit
Wendy Reid: what we see as problems for EPUB/A are external links (images, video, audio, etc.)
… if a publisher creates an EPUB without external resources should be ok for EPUB/A
Ivan Herman: what you are defining are general guidelines on how to produce EPUB files to be archived
… and not a new format, is it correct?
alicia: it'll be a subset of an EPUB
Ivan Herman: I understand, but there is a difference between defining a list of features you can use to create an EPUB/A, that is different of taking the spec a remove contents
… the second option is more difficult, because the docs are structured differently
… from 3.0.1, it would me more simpler if you list the features to avoid in the EPUBs
Tzviya Siegman: with wendy we started the discussion, for what we've seen the document should be quite different form a spec, it'll be compatible with EPUB 3.0.1 and EPUB 3.3
alicia: your group is an important stakeholder in this process
liisamk_: we've published the first ebook 25 years ago (before EPUB), we've republished it multiple times; it's a software: we update it
… from an archival point of view, do you pick these publications multiple times? do you take the first? do you take the latest?
alicia: it really depends from archive to archive (based on national legislation)
<Murata Makoto> I can explain what JWG7 is. I have been involved in it from the beginning.
Brady Duga: please don't call "subset", I think is better "EPUB with additional requirements"
<Wendy Reid> +1 Brady Duga
Bill Kasdorf: I understand that you're creating a format for a "dark" archive, that is not public
… that is not available to libraries
alicia: the content that has been deposit is inaccessible to everyone
… this is for preserving content for the future
<Tzviya Siegman> +1 to a11y metadata in archives
George Kerscher: EPUB 3.3 includes accessibility, I think we have the possibility here for the first time to include accessibility metadata in archived documents
alicia: what kind of accessibility metadata?
George Kerscher: these metadata are not required by EPUB 3.3, but in EPUB Accessibility 1.1 requires acccessibility metadata in the publications (e.g. conforms to WCAG, etc.)
… we're seeing retailers displaying these metadata
Avneesh Singh: EPUB Accessibility applies constraints to EPUB, so that tells how to make an EPUB accessible
… it applies to different versions (EPUB 2, EPUB 3.0.1, EPUB 3.3)
… we'll be the EPUB/A written in the same way?
here is the ISO spec for the EPUB Accessibility 1.0 - https://
<wolfgang> s/writtern/written/
<Ivan Herman> qq
alicia: I think ISO we'll be based on the latests version available as ISO
Murata Makoto: I've been involved in JWG7 WG for many years
Most recent EPUB Accessibility 1.1
Conformance and Discoverability Requirements for EPUB publications https://
Murata Makoto: when IDPF developed EPUB 3, they decided to spread the format through ISO
… using different standards (please Makoto fill in the numbers :) )
… I think that a lot of other people are interested in EPUB/A
… this impacts other formats like SGML, XML and OPF
… all these are maintained in ISO
… tomorrow we'll have a WG meeting to discuss this in ISO, with EPUB 3.3, EPUB RS 3.3, etc.
Gregorio Pellegrino: One from the side of the industry, it might be difficult or too expensive to make 2 epubs
alicia: No intent to force both versions
Ivan Herman: Alicia you said "we'll refer to any EPUB version is recognized by ISO"
… creating a document that will fit to EPUB 3.3 may not be that simple
… I don't know if you have something similar for websites
alicia: in the document we've requirements for avoiding scripting in EPUBs
… but we're doing R&D on this
Tzviya Siegman: thank you alicia
can someone else take the scribing? :)
Tzviya Siegman: we know that there are concerns for taking EPUB 3.3 to ISO for EAA
<Tzviya Siegman> scribe?
EPUB to ISO
two cañas for Brady Duga
Tzviya Siegman: Pro and anti ISO move forward
Murata Makoto: Korea is an important player here, they have (???) and 3.01 as national standards
… (3.0)
… This was based on an agreement with IDPF
… They are concerned and some joint WGs will be discussing this shortly
… the day after tomorrow
… And we hope they will send a statement to pub @ w3c
… I think it is important we listen to Korea
Wendy Reid: To contextualize, this is less pro and anti
… It is more a timing problem
… hopefully that problem will go away
… We want to have the best standard as we approach EAA
… It is really important the EAA move ahead
… But there are also many countries outside the EU that rely on ISO as the standards body to define things
… We want to put this decision to bed today so we have a plan
… the question isn't if we will send this to ISO, it is when
… Currently we are stuck in a grey area where we don't know dates and timelines
Shinya Takami: Who is the Korean that asked for this?
Murata Makoto: (some name, I didn't catch)
Shinya Takami: Is he a member of w3c?
Murata Makoto: No
Tzviya Siegman: It would be nice to have them present so we can discuss
Murata Makoto: It might be good to have a joint session between JWG7 and us
<Murata Makoto> 3.0: TS; 3.0.1: IS
George Kerscher: 3.0.1 was submitted as a technical spec. In terms of signed agreements, we only required that the spec would be public (no payment required)
… at least that is my recollection
Murata Makoto: 3.0 became ts, 3.0.1 became an international standard
… both were fast tracked by Korea
Charles LaPierre: As far a timing, I assume we have to accept all the outstanding PRs
… and make sure that there are no outstanding issues
Ivan Herman: Yes
<Murata Makoto> Agree with <Brady Duga>
Ivan Herman: 3.0.1 is an international standard
… it is very out of date
… we must use the w3c process to get to ISO
… Do we or do we not have information that publishing epub 3.3 as ISO standard is harmful to EAA
… My problem is the EAA process, it is all the bothers me
<Avneesh Singh> +1 main question is EAA
Murata Makoto: I would like to point out outdated ISO standard is harmful to countries that rely on it
Murata Makoto: WCAG 2.1 is unusable in Japan, I don't want that to happen for epub
Christina: It is not just Europe
… Publishers can work now on a11y without further requirement
… In my opinion is wait until the law is in effect, June 2025
… Changing the standards now will create a lot of confusion in the industry as to which one to adhere to
<shadi> +1 to Cristina
Christina: the commission can accept the w3c standard
… If it is not absolutely mandatory, then we should wait for EAA is in place to avoid confusion
Ivan Herman: The date for EAA is beyond our charter
… The 3.3 standard will remain a w3c standard for the foreseeable future
… The document isn't transferred to ISO
… Is this differentiation meaningful to the EU?
… If they suddenly disregard the w3c standard because of the ISO, it seems stupid,
Wendy Reid: Agree with Ivan Herman
… It seems unfair to countries to say "wait 2 years to start"
… It impacts people like Alicia and others, and seems unfair
… Need to find a sensible resolution between EAA and other people
Avneesh Singh: Want to highlight there are epub 3.3 and a a11y 1.1
… The bottom line is the main reason epub a11y 1.1 exists is to support EAA
shadi: 2025 is the furthest out
… If ISO standard exists it means the commission cannot reference a w3c standard
… If an ISO standard exists it *must* be referenced
… the best way forward is to have consensus
<Murata Makoto> Sounds great to me.
shadi: We should invite the EAA as proposed
… the process is relatively simple, it is possible for the commision to issue delegated acts
… So when are we ready to say he is a11y 1.1, when will FXL be covered, then put out that timeline
… The commission will not act by itself, it needs us to push forward
Tzviya Siegman: I don't understand
Tzviya Siegman: what is the timeline?
… is this just an intent to move forward?
shadi: That is just an input to the commission to move forward ASAP
… We have a good case to have them expedite the process.
… June 2025 is when they start policing. We have a good case for them to move forward quickly
Tzviya Siegman: This is more than just a11y
Ivan Herman: I am trying to understand what we need to do
… 1. need to finalize the document for fxl a11y issues
… That is the first step, need to publish as a note
… Maybe 2 months?
Wendy Reid: [laughs]
Ivan Herman: Then we go to the EU with 3 recs, plus 2 notes.
… This should be considered the oracle for what we need
… If we get the nod from EAA then we can move forward with ISO
Ivan Herman: Is that the correct timeline
shadi: It doesn't have to be that sequential
… we can already start
… We can start the process for a delegated act now
… even without the fxl doc
… We don't have to wait until the note is done to talk to the EAA
Ivan Herman: It seems like we are all waiting for each other
christina: We can start with the mapping, and also say that we are also working on the metadata display, etc
… Start with what we have, then add fxl
… Since that is one case of what is already in the mapping
<shadi> +1 to Cristina
Wendy Reid: We will discuss at a11y tomorrow
… We, as a taskforce, did not know the gravity of the fxl note
… Can we do it in a reasonable amount of time? Probably, but we need to look at the schedule for that
Tzviya Siegman: I think given the state of fxl, is it possible to submit without it? There isn't even a draft
several: there is!
Ivan Herman: Closing the discussion, is there general agreement in the group to publish as ISO providing it doesn't jeopardize the EAA
<Murata Makoto> Sounds good to me.
Ivan Herman: We should have an agreement on whether we will ever move forward with ISO
Avneesh Singh: Regarding the discussion with EAA, the starting point is the mapping
… from there we can expand to other documents
… we can start there, then add fxl
George Kerscher: My understanding is that if we can get the delegated doc (delegates to our a11y spec), then we can move ahead with ISO
… and it won't jeopardize Europe
shadi: Yes, once we have the delegated doc we can move forward with ISO
Ivan Herman: We need a formal resolution
<Tzviya Siegman> proposal: WG agrees to publish the 3 recs as ISO standard under PAS process once the European Process is completed
<Tzviya Siegman> +1
<Shinya Takami> -1
<Wendy Reid> +1
<Ivan Herman> +1
<shadi> +1
<Bill Kasdorf> +1
<liisamk_> +1
<wolfgang> +1
+1
<Laurent Le Meur> +1
<Murata Makoto> +1
<Masakazu Kitahara> +1
Wendy Reid: Shinya, do you formally object to ISO?
Shinya Takami: Yes. We do not need it
<Masakazu Kitahara> Sorry. 0
Shinya Takami: I don't think there is a negative impact if we do not have it
<Toshiaki Koike> +1
Wendy Reid: Do you really formally object, or can you live with it?
0
<Murata Makoto> The JP government has not adopted WCAG 2.1.
Shinya Takami: There is a problem with moving a11y 1.1 to ISO, which has a negative impact
Wendy Reid: 1.1 adopts the current version wcag
<George Kerscher> +1