EPUB 3 Working Group Telco — Minutes

Date: 2021-09-10

See also the Agenda and the IRC Log


Present: Ivan Herman, Murata Makoto, Dave Cramer, Jen Goulden, Avneesh Singh, Deborah Kaplan, Wendy Reid, Toshiaki Koike, Gregorio Pellegrino, Masakazu Kitahara, Ken Jones, Laura Brady, Tzviya Siegman, Charles LaPierre, Dan Lazin, George Kerscher, Brady Duga, Ben Schroeter

Regrets: Matthew Chan


Chair: Dave Cramer

Scribe(s): Gregorio Pellegrino


Dave Cramer: today we’ll not work on issues, but more on feedbacks from task forces and TPAC
… first thing is an update on testing

1. Initial testing instructions (pr epub-tests#23)

See github pull request epub-tests#23.

Dan Lazin: I’ve prepared a document on testing instructions
… then we have to put metadata so we can create automatic reports from the metadata
… so we can invite non-testers to start testing
… we are looking for volunteers for creating tests for FXLs and MOs EPUBs

Ivan Herman: I have a system for automatic reporting but only in a test repository, I will add it to the tests’ repository when it is final in its structure
… this system will create reports automatically for every new implementer

Dave Cramer: can you describe how the reports will work?

Ivan Herman: every implementer should submit a JSON file (super simple structure) giving a true or false if the test worked or not
… then the system will get all the results and create the report
… it’s really similar to the one used for audiobooks

Brady Duga: what it means that a reading systems pass a test?
… I have three reading systems, what happen if one passes and the others not?

Ivan Herman: It’s up to the WG or to the implementer to decide if it is a unique reading system or that are three different reading systems
… the goal of the testing is not a beauty contest, the goal is to check that the features in the spec are consistent and can be implemented

Dave Cramer: from my point of view, the google play books is one system (with three implementations)

Brady Duga: we have three different teams, the clients are different, but the backend and the pipeline is common for all three
… so maybe yes, it’s only one system
… but what does happen if one implementation pass and the other two doesn’t?

George Kerscher: for accessibility we differentiate the OSs, since AT + OS + reading system may have really different results

Wendy Reid: I think that in the instructions we should say that depending on how you have implemented you systems, it may be worth to test the different apps on different OSs

Brady Duga: +1 to wendyreid

Tzviya Siegman: I think the epubtest.org is super important to understand how reading systems work on different OSs
… it’s similar to ARIA: ARIA is not tested on the browser, but on the Accessibility API exposed by the browser engine
… I don’t know if it is possible with EPUBs

Tzviya Siegman: https://w3c.github.io/test-results/wai-aria/all.html ARIA Tests

Brady Duga: I don’t know if it’s meaningful: some work is managed by the backend, other is managed by the frontend

Ivan Herman: I propose to report results in two formats: one for the W3C (unique implementations), the other for the community (separating the OSs). This will require additional metadata in the implementation report, but it should be doable

Brady Duga: +1 to ivan

2. planning for TPAC

Dave Cramer: I think we have some important issues to discuss on URI-s, but I think we need people from reading systems and from EPUBcheck to discuss with them. I am thinking of :

Dave Cramer:https://github.com/w3c/epub-specs/issues/1681 (root-relative paths)

Dave Cramer:https://github.com/w3c/epub-specs/issues/1374 (IRI of package document)

Dave Cramer:https://github.com/w3c/epub-specs/issues/1686 (duplicate manifest entries)

Dave Cramer: we also have some issues tagged as accessibility
… or ask the accessibility TF to take care of them

Avneesh Singh: not all may be related to only accessibility

Avneesh Singh: some issues require broader discussion
… I will go through the issues and inform chairs

Dave Cramer: we have a label on github for tagging issues that we want to discuss during TPAC
… so the accessibility TF may tag those issues

Ivan Herman: there are some issues on vocabulary

Tzviya Siegman:https://github.com/w3c/epub-specs/issues/1763 Should all Structural Vocabulary Items be normative?

Dave Cramer: yes, we have to check with publishers if those terms are used in EPUBs

Tzviya Siegman: we may ask to distributors to check programmatically check EPUBs
… to get statistics

Ivan Herman: yes, this approach may be important in the process of making terms normative
… if it’s possible to have from two different sources it will be great
… (Google, Kobo, etc.)

Brady Duga: I can create a report, but I need time..
… beginning of October

Wendy Reid: is it about epub:type vocabulary?

Ivan Herman: any vocabulary that is in the core spec

Dave Cramer: we should document what we are asking to platforms
… Task forces

2.1. fxl accessibility

Wendy Reid: FXL accessibility TF is going well
… we are writing the documentation, there are some sections we have to complete
… we haven’t started to map content to WCAG
… we are finalizing the proposal for switching from textual to visual

2.2. virtual locators

Wendy Reid: Virtual locators TF: is going well, we are working offline
… we spent a lot of time working on terminology
… now we’ll work on proposals for solutions
… probably those will work very well for textual books, but maybe not for books with images or formulas

2.3. a11y task force

Avneesh Singh: on accessibility TF: the major objectives are done, one issue there is bigger
… on conformance, checking if zero tolerance is practical
… we’re not sure if we can fix in this version
… for back-compatibility

Avneesh Singh: See A11y related issues

Avneesh Singh: for TPAC we’ll defining meetings with APA
… we are fixing a meeting on accessibility metadata
… one of the point on the agenda with APA is the accessibility horizontal review
… but we should trigger the formal request for the review
… may we move on?

Ivan Herman: there are also some accessibility metadata that refer to schema.org
… but right now we don’t know if we can refer to those terms as schema.org’s terms
… I started a discussion with the director to check what we can do about it
… one solution may be to create a CG for maintaining this vocabulary
… Matt as already prepared a PR on this

Avneesh Singh: maybe a task force inside schema.org CG is better than creating a new CG

Tzviya Siegman: is this an addition to schema.org?

Charles LaPierre: our values for the schema.org properties

Ivan Herman: no, the fact is that relevant terms in schema.org are not very stable, so we cannot refer to them normatively

3. AOB

Dave Cramer: on review: I’ve sent a questionnaire about security and privacy, the PING WG will have a meeting with this on their agenda next week.
… next week me and Wendy will be at the meeting

Charles LaPierre: APA horizontal review?

Ivan Herman: please, Avneesh, trigger the APA review

Avneesh Singh: I’ll do