EPUB 3 Working Group Telco — Minutes

Date: 2021-01-21

See also the Agenda and the IRC Log


Present: Toshiaki Koike, Teenya Franklin, Zheng Xu (徐征), Matthew Chan, Wendy Reid, Ben Schroeter, Marisa DeMeglio, Brady Duga, Masakazu Kitahara, Ryo Kuroda, Garth Conboy, Shinya Takami (高見真也)



Chair: Wendy Reid

Scribe(s): Matthew Chan


Wendy Reid: welcome everyone. 2 items on the agenda today

Wendy Reid: https://github.com/w3c/epub-specs/wiki/Big-Issues-for-EPUB-3.3

1. Scope

Wendy Reid: this is a list of various categories of issues we have in epub 3.3
… some of these we’ve already discussed to some extent
… we’re coming to a point where we have to determine whether each of these issues meets the test of being within our scope
… e.g. HTML serialization, should this be in 3.3? Or is it something that should be left to a later time due to the risk it poses to backwards compatibility?

Garth Conboy: is there a particular issue that we want to decide on right now?

Wendy Reid: no particularly, we want to understand what is high on everyone’s priority list
… which ones are important to the community that you represent, or which ones do we need to get more information about before we can decide?
… we also want to be up front about issues that we have to move out
… especially as we run into issues with testing, or backwards compatibility

Ben Schroeter: can we run through these at a high level? i.e. a summary of each?

Wendy Reid: html 5 serialization, essentially bringing html5 into the spec officially, making it live in parallel to xhtml
… core media types, which new types do we add? how far do we take it?
… fate of epub:type. We can’t fully remove it, but what could we do about potentially deprecating it, or steering people away from it
… fate of multiple renditions, we’ve pretty much decided
… fate of satellite specs
… inline formatting in nav doc
… scripting, we’ve already discussed to some extent
… origins and remote content, are somewhat related
… complex layout, I believe this has to do with mixed modalities, no pixel values in complex layout content
… fate of linked metadata records, something dave has been trying to test
… fate of manifest fallback, not shown to be highly supported in testing, but which still exists

Ben Schroeter: nothing jumps out as conceptually out of scope, maybe out of scope due to lack of bandwidth

Wendy Reid: don’t want to write anything out, but, e.g. linked metadata records and manifest fallback are not well supported, should these be lower priority?

Ben Schroeter: do those satellite specs fall under our auspices?

Wendy Reid: yes, they are referenced in our charter, but a lot of them aren’t note status (except for multiple rendition)
… a lot of them are currently IDPF documents
… we just want to get everything thinking in terms of this list of issues

Garth Conboy: i was curious if there was anyone on this call that hasn’t had a chance to express opinions on these issues
… especially on HTML serialization

Wendy Reid: we’re still gathering information on that topic
… particularly, we don’t know whether Amazon and Apple have plans to support HTML5, or a desire to

Teenya Franklin: I am literally a beginner (in fact a complete beginner) on Epub and web development and honestly while I learn alot on these calls I do not know enough to express and opinion. I do however appreciate the opportunity to be a member and learn as we go.

2. Authoring and HTML5

Wendy Reid: Authoring tools includes InDesign, Sigil, or even a completely custom publishing workflow that someone has developed for their indie publisher
… we need to consider what challenge HTML5 would present to people making the tools in the tool chain

Marisa DeMeglio: if you have authoring tools creating xhtml5 currently, does anything get lost if we just convert that to html5 after the fact?

Wendy Reid: i would guess no…

Ben Schroeter: epub:type?

Wendy Reid: might be affected, but we want to discourage them from doing so anyway
… someone commented on the original thread from Standard Ebooks. They convert public domain books into ebook files
… they don’t have custom tools, but they pull from standard libraries to create a tool chain
… this person was reluctant to move towards html5 serialization because of what would be lost vs xhtml

Wendy Reid: also, xhtml already contains a lot of the good things from html5, so do we even need the full thing?

Toshiaki Koike: comments in japanese

Shinya Takami (高見真也): Toshiaki Koike says he is making generation tools for epub content, for which he currently uses an xml parser
… if we changed to html5, this would require big changes in his process
… other publishers or vendors may be similarly impacted

Zheng Xu (徐征): isn’t the proposal to ADD support for html5? xhtml would still be accepted anyway
… so this might limit the degree of the impact of html5 serialization

Garth Conboy: Google uses an XML workflow, and I would expect others do as well for historical reasons
… but that in itself is not a reason to say no
… and a change in this direction might be helpful for other segments of tooling that are based on html5, and also for scripting
… it would be nice if we could get input from the business group as to who wants this, and how badly?

Wendy Reid: we have heard from independent publishers, as opposed to the people who make the tools that are being used in publishing
… that was where the support for html5 serialization mostly came from
… but we do need to have a better understanding of how existing tools would be affected, how much would need to be changed. Is it just adding an html5 parser? Or it would be more work?

Ben Schroeter: are we saying that we want future epubs to be strictly html5 with no extensible markup?

Wendy Reid: no, we want to take advantage of html5 flexibility and modernity
… we’re not expecting html5 files to in every epub.
… but if we allow it, we’ll probably see some of both for a very long time

Garth Conboy: there is a formality to xhtml that is useful, as opposed to the tag soup that is html serialization

Ben Schroeter: do we need more of in impact investigation? Input from RSes, bookstores, etc.?
… we don’t want to make assumptions and miss anything critical

Wendy Reid: we’ve already asked RSes, pretty much, but we need input from tool makers
… e.g., the makers of pandoc
… also the parts of the publishing industry who haven’t provided input (e.g. apple, amazon)
… if you are in contact with people in these groups, please ask them for feedback
… so, basically, more research before we decide

Garth Conboy: one of us could probably write an email to Tess, just to ask for Apple’s input

Wendy Reid: yes, we’re going to do that

3. AOB?

Wendy Reid: we published FPWD just over a week ago
… it was well received
… but we already have a number of changes since then
… we’re not ready to publish anything else yet, but as a protocol, we’re going to resolve it in a meeting first
… as opposed to any sort of automatic publishing
… so, yes, we want your input as to when we publish
… the next major publication is probably going to be aligned with the accessibility piece
… we’ll see you all next week everyone!