Meeting minutes
Restarting chartering
<Wendy Reid> https://
Wendy Reid: Announcement - at the end of this chartered I will be stepping down as chair
… Do not fear! We have a replacement
… Brady Duga will join as chair
… Susan Neuhaus and shiestyle will be continuing
Ivan Herman: administrative - Brady Duga has joined the chairs calls
… and Wendy Reid will likely stay on for a bit
Wendy Reid: We will be publish a start of the charter, please review it
… will include next epub version, annotations, etc
Ivan Herman: There is a link to a diff of this and the previous charter
… the second section, Scope, will need discussion
… the big yellow blob there will need to be worked out in Dublin
… These are some things I thought of, plus some items from our previous recharter discussions
… The important point is that EPUB annotations will not make this cycle, it will be under the new charter
… I have made some entries about the ISO work. I made that a new section, the old charter talked about it in a vague way.
… it is better spelled out here
… after those sections it is mostly admin stuff
… 3.4 is the timeline, comments there are welcome
… Timelines are a rough indication, we rarely consider them after chartering them
… Before Dublin, in the liason section, I think there is a new multimedia TF in the CG, can I get contact info for that?
Laurent Le Meur: At TPAC it looks like the charter is at the end of the day, can it be moved up?
… I have a flight out on Tuesday afternoon, I would like to participate in the discussion
Wendy Reid: The agenda is still fluid, we can make it work
Dale Rogers: I notice audiobooks is listed under our umbrella
… we never talk about them, is it because they are really stable
Wendy Reid: We developed a spec, it is used, but not widely
… it is also very fluid
… we keep it there if people want to add a bug, etc
Brady Duga: Playbooks supported the audio format, and publishers said they wouldn't use it any more, but use the Kobo format
Wendy Reid: the formats are very close
Hadrien Gardeur: Audiobooks spec can succeed for 2 reasons
… there is interchange, and there is end user adoption
… spec has failed as both
… I have never seen it used as end user
… For interchange, not really used and lots of people are asking for different things
… and that can vary by country
… so the workflow is whatever they are being asked for
… I don't think there is much leverage for standardizing for interchange
… the problems aren't really technical, we would need something external (eg a trade org) to push the format
… Real change can't come from us, needs to come from one of those orgs
… For end user, it really isn't needed. DRM free is easy, and it doesn't help with DRM, so not really useful
Ivan Herman: This came up because of the charter, and that doesn't need changing
… it doesn't need to be changed in the charter
… There is a mechanism to drop a spec (rescind), but we are not discussing that. so the charter is fine
Annotations issue #3009 - w3c/epub-specs#3009
Wendy Reid: This is about metadata in annotation sets
… we discussed whether the metadata in the epub is good enough to include, but we ended up that it can be useful, eg for an end user to identify the book
… so what metadata still needs discussion
Ivan Herman: We should never look at these spec in isolation. It shouldn't be the job of the annotations spec to go into discussion of the metadata or the quality of it
… Whatever the package metadata quality is, it is what we have
… if there are problems with the metadata, we should discuss it as part of EPUB, not annotations
<George Kerscher> +1 to what Ivan Herman said.
Ivan Herman: We should preserve the package MD as is
Laurent Le Meur: I agree. I have a PR where I pulled dc:format. Since we only discuss epub it is meaningless
… In spec we say the metadata can help with matching, but it is very vague
… dc:publisher doesn't seem to be described in 3.4
… publisher is a creator in 3.4
Ivan Herman: If it isn't in epub, we should remove it. And the text should say this is a copy of what is in the publication
… it isn't our job to make a better identifier, that should be addressed by epub 3.5
Laurent Le Meur: I made a mistake, dc:publisher is there
Hadrien Gardeur: I really think dc:identifier is not the average metadata
… it isn't our job to redefine it
… it is a trap
Brady Duga: I agree I understand the desire to copy the md that is in the epub. But this dc identifier means nothing to the user
… they know about the author, title, publisher. Identifier is useless. It really is a different piece of information from all the other information
… if we include it, we should be clear not to show it to the user. And that is confusing
Ivan Herman: I have no problem not to include it
Brady Duga: we should add a warning if we include dc:identifier that you shouldn't use it to identify books or show it to readers
Ivan Herman: 2 questions - one question is what to copy here
… other question is should we re-open dc:identifier discussions in Dublin
Wendy Reid: we did also discuss generator, we did have issues about PII
… so we leaned against that - what about other parts?
George Kerscher: Assuming the annotation set is kept after loosing the book, I would like to see us add whatever is needed for a proper citation
… and that solves the identification issue
Wendy Reid: The list in the About object has: identifier, format, publisher, and date
… what else do we need
… ?
George Kerscher: I don't know if that is enough - date is one problem
… Publisher maintain a list of what is needed
Susan Neuhaus: We decided to drop format
… Right?
Wendy Reid: Does that matter?
George Kerscher: Yes, lots of accessibility things
<Hadrien Gardeur> We don't put EPUB version in the media type
Wendy Reid: It's just a media type, no version
Charles LaPierre: Don't we need page numbers?
Laurent Le Meur: Yes, but that is at annotation level, not set
Ivan Herman: There are some formats that need eg city of publication
… I come back to the same problem, all we have at annotation creation time is what we have in epub
… the format is of the citation is not part of this interchange format
Brady Duga: I agree with Ivan Herman, there is a question of citations, and that is not an issue for annotations
… perhaps we could add citations to the new charter, and then update the annotations
<Charles LaPierre> +1 to Brady Duga
Brady Duga: but this shouldn't be part of the annotations discussion now
<George Kerscher> +1 to add what is needed for citations to the charter.
Wendy Reid: Ebook citations are different than regular
… for ebooks there needs to be a stable url or identifier
… It is often a link to a products page
… these are rarely part of the epub
Hadrien Gardeur: I understand the need, that said I don't think we should attempt it
… I don't think we can do it from the EPUB side
… this only works in a closed environement and getting metadata from elsewhere (eg ONIX)
… In that case the RS may have all the data it needs
… the other use case is a silod file
… If you are an open RS the only thing you can do is grab the data you can find, and let the user fix it
… I don't think we can do better than that
… From a reading app perspective, those are our options
George Kerscher: in higher ed, I get this request all the time
… Students start with the epub, and then get stuck trying to make the citation
… we have to address it
Laurent Le Meur: There is another use case, where there can be a different app (eg Zotero) that can ingest the annotation set and create the citations
… today the metadata is poor
… but I have taken what we have
Brady Duga: this is a specific bug in citations, and a separate topic from annotations
Ivan Herman: I will add a line to the yellow blob about metadata for citations
… maybe we won't be able to do anything, but we can decide in Dublin
… we could add a line that says it is fine to add your own metadata
Laurent Le Meur: There is already a sentence about that
… which allows other vocabularies
Wendy Reid: Laurent Le Meur do you have enough info?
Laurent Le Meur: Yes, I will add this to the PR
Annotations issue 3037 - w3c/epub-specs#3037
Brady Duga: right now, the target of an annotation is any publication resource
… you could bookmark and image even if it is a spine item
… so I propose the target be only a spine item, and not any resource. You could reference a file that the user never even sees
… CFI natively supports this
… Laurent Le Meur wrote some text about this, but we never have defined spine item, even though we use the term alot
Hadrien Gardeur: I agree, since eg images can be repeated
… Images can be referenced multiple times with different alt text, etc
… I would add a warning box for non-linear
… which seems to be an additional risk
George Kerscher: We use non-linear for extended descriptions at the end of the book
… I imagine that someone might want to annotate the extended description
Susan Neuhaus: Footnotes are also often non linear
Hadrien Gardeur: Just to clarify, I don't say we should drop them, just that you are living dangerously when you use them
Brady Duga: I see your point, it is risky, I don't know what the rs exporting the annotations can do about these risks
… what would Playbooks do, for instance? Warn the user?
Hadrien Gardeur: The risk may be around progression, so that can be incompatible
… it is additional risk, so things that make sense for linear may not make sense for non-linear. I think it is worth pointing out
Ivan Herman: is it a problem with the annotation in general, or with specific selectors
Hadrien Gardeur: Neither, the core issue is how non-linear is handled
… So the inconsistent implementation is the problem
George Kerscher: In the case of extended desc, the original image is always referenced
… could we make that association? Or will we force people back to the image
Brady Duga: I don't think we're saying that you won't be able to annotate an extended description, but you'll have to be careful how it is exported
… I don't think we can require that you go back to the original image
… are we somewhat agreed that there will be some restriction to spine items, maybe just warning?
… we should define spine item in the spec, and then link to that in the annotation spec
Ivan Herman: I think we can do it in 3.4
… it is not a fundamental change
Laurent Le Meur: Ok, I will just reference spine item and wait for the spec change