Meeting minutes
<graham> Regrets from Rick, Dahei
graham: first agenda item is report from publishing maintenance working group
<ivan> first day
<ivan> second day
shiestyle: we met last week in Sofia, face to face. we discussed webtoons, annotations and other topics. About 15 persons attended.
graham: is there a time to expect available?
shiestyle: we hope to have a draft next year.
ivan: the charter is for 2 years, there are few items to add to the spec, but it also always take more time than expected.
gautierchomel: we wrote an article to resume : https://
ivan: two interesting questions appeared to me, the one is the use we make of dark mode, how it could combine with themes on reading system is not a trivial question.
ivan: the other is HTML in EPUB, we know we need to keep XHTML long term support, and we know it is difficult to move to HTML. We'll allow HTML, it does not mean you are required to use it. We hope implementation feedback to understand the effects.
Bill_Kasdorf: does it means that reading system should be neutral on supporting both HTML and XHTML?
ivan: yes. That's does not seems to be a big issue, they already rely on browser that support HTML.
graham: next item, Business working group report
liisamk: we had meeting, we encourage participation on TDM webinars.
liisamk: also had a presentation from orange, the tools and methodology they have to deliver faster translation for mangas using AI and humans. They say it is a way to fight piracy (because legal offer exist).
liisamk: we'll have more discussions about AI business case.
graham: last year Editeur proposed a paper about how the use of AI should be informed in metadatas
<graham> AI in ONIX - https://
graham: couple of aspects are covered, publisher used AI in the creation process means AI has to be informed as contributor; the TDM rep protocole that must be duplicated in ONIX and content.
Crstina: to me using AI to translate quickly was an unexpected way to fight piracy, very interesting.
graham: it is he same process to fight piracy existed in paper, years ago. It is about having a legal offer, even if at low cost, to exist in the market.
ivan: we start to see converging interest in W3C worlink group about AI, it's worst keeping an eye on different works.
ivan: on annotation, it would be good to have a business working group looking at the work we start, even great if someone can participate. We need to make sure there is an identified business case. A practical approach.
graham: on AI we encourage developers of metadata process management system to quickly add those information in theire system so publishers can use them.
<graham> re AI,( encouragement to use AI for creating alt descriptions for accessibility purposes)
gautierchomel: is there a metadata to inform the use of AI to generate alt text?
graham: yes, we have a way to credit alt text, meaning if it's AI generated, it could be creadited this way.
liisamk: it's usually melted creation, a bit of machine, human review. It is difficult to give an accurate credit here.
Bill_Kasdorf: there are also companies that provide alt text creation service, we don't always know clearly how they work
graham: next topic, community group report
wolfgang: AI for us too. We had a presentation of the use Vital Source makes of AI to enhance learning path by creating quizzs of a textbook. It's AI, but not LLM yet. They are reticent because they want to make sure it's a safe environement both for published data and for student using it. It was interesting the way they use AI, not only practically but also ethically.
wolfgang: we'll follow up this series of AI in reading systems. We expect to have a presentation of Chaï App, Colibrio, Thorium and any more we find on the way.
graham: next topic, accessibility task force.
George_: the PCG A11Y TF published the display accessibility metadata guideline, we had to do more importnat changes but we are very close to a final. There is a JSON file to be localised, Vital Sources offered 30+ translations. We have a viewer to test.
Display guide draft : https://
graham: wonderful work that will help us following advocacy with publishers.
gautierchomel: there is a Accessibility metadata display processor
https://
gautierchomel: it's just been implemented in readium mobile toolkit, soon thorium desktop and web will follow.
liisamk: are we tracking how that will be used by reading systems?
Cristina: we are implementing it on the LIA catalog.
graham: not sure what we can do here. I would be happy to find a strategy.
Bill_Kasdorf: I would like to add to next meeting agenda a discussion on epubcheck version number strategy. The fewer changes the better, so high frequency release is easier for developers and maintenance. We suggest two release a year, at fixed dates, end of summer and early January, one being only maintenance.
ivan: we also have to think about how we plan evolution of epubcheck in relation with EPUB3.4. I think about HTML obviously. It's a complicated part to implement in a checker.
graham: ok, adding this item to next meeting.
George_: AB elections will happen on W3C, Avneesh is candidate, I invite each here to vote for him.