search

w3c / publ-cg

EPUB 3 Community Group Repository
Other
44 stars 16 forks source link

Maintaining EPUB 3.X: Versioning, Dating, Living Standards, etc. #97

Open dauwhe opened 5 years ago

dauwhe commented 5 years ago

The very mention of the phrase "living standard" has raised some eyebrows. How best can we maintain EPUB 3.X? Is it helpful to anyone to keep creating new versions of EPUB? And, most importantly, can we take any steps closer to our original goal, where the spec represents a contract between the content author and the reading system, an assurance that if your content follows the spec the reading system will render according to the spec.

Let's look at an example. The biggest recent question has been about spine order vs nav order. What happens if we decide to relax this restriction in the spec? Do we have to publish EPUB 3.2.1 or something, even though the change will not affect any existing EPUB or any reading system?

Actual EPUBs are not strictly tied to versions. Most of what my employer produces probably validates as EPUB 3.0, EPUB 3.0.1, and EPUB 3.2. But what if I embed a WOFF 2.0 font in my EPUB? That would be invalid in 3.0.1 but valid in 3.2. That still doesn't tell us if the underlying rendering engine supports WOFF 2. CSS, of course, is designed to fall back gracefully if it sees a font format it doesn't understand. MathML is another example. Can I actually use it in my EPUB? Maybe.

To be continued…

artbyrt commented 5 years ago

Current versions of epubcheck flag WOFF 2 with an informational warning; curious as to why that would be if it’s part of core epub 3.01-3.2.

Also: If EPUB 3 spec is subsumed in W3C, doesn’t this automatically mean–at some level–measured improvements in the spec and thus the rendering engines? For example: ‘footnotes’ and the issue of a ‘footnote’ tag?

ruth tait

On Aug 7, 2019, at 10:57 AM, Dave Cramer notifications@github.com wrote:

The very mention of the phrase "living standard" has raised some eyebrows. How best can we maintain EPUB 3.X? Is it helpful to anyone to keep creating new versions of EPUB? And, most importantly, can we take any steps closer to our original goal, where the spec represents a contract between the content author and the reading system, an assurance that if your content follows the spec the reading system will render according to the spec.

Let's look at an example. The biggest recent question has been about spine order vs nav order https://github.com/w3c/publ-epub-revision/issues/1283. What happens if we decide to relax this restriction in the spec? Do we have to publish EPUB 3.2.1 or something, even though the change will not affect any existing EPUB or any reading system?

Actual EPUBs are not strictly tied to versions. Most of what my employer produces probably validates as EPUB 3.0, EPUB 3.0.1, and EPUB 3.2. But what if I embed a WOFF 2.0 font in my EPUB? That would be invalid in 3.0.1 but valid in 3.2. That still doesn't tell us if the underlying rendering engine supports WOFF 2. CSS, of course, is designed to fall back gracefully if it sees a font format it doesn't understand. MathML is another example. Can I actually use it in my EPUB? Maybe.

To be continued…

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/w3c/publ-cg/issues/97?email_source=notifications&email_token=ADGANABA5LU65VIL2IL3NW3QDLPFNA5CNFSM4IKBK2D2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HD5NGJA, or mute the thread https://github.com/notifications/unsubscribe-auth/ADGANACBDAD3B43ZPWIONO3QDLPFNANCNFSM4IKBK2DQ.

laudrain commented 5 years ago

Before making the spec evolve, shouldn’t we look at implementations of EPUB 3.2 to raise the quality of global adoption on the Reading Systems side?

There is a lot to do, like for instance in accessibility where “born accessible EPUBs” are not rendered properly to VIP.

Luc

dauwhe commented 5 years ago

There is a lot to do, like for instance in accessibility where “born accessible EPUBs” are not rendered properly to VIP.

What does "VIP" mean in this context?

mattgarrish commented 5 years ago

VIP = visually impaired persons.

On a practical level, in a world with dozens and dozens of "reading systems", how many implementations do you need before you can really consider a feature supported? What market share qualifies? In whose region?

The one concern I have with a "living" standard is that without revisions the decision making process may get watered down. There's always been a noticeable uptick in participation in the past when we've focused people on completing a revision and had a set timeline to get things done. If we just expect people to stay constantly tuned in to every change we likely won't get the same attention to what's going on.