Publishing Working Group Telco — Minutes

Date: 2020-02-03

See also the Agenda and the IRC Log

Attendees

Present: Laurent Le Meur, Ric Wright, Dave Cramer, Ivan Herman, Wendy Reid, George Kerscher, Franco Alvarado, Gregorio Pellegrino, Tim Cole, Avneesh Singh, Bill Kasdorf, Ben Schroeter, Charles LaPierre, Tzviya Siegman

Regrets: Garth Conboy, Geoff Jukes, Marisa deMeglio

Guests:

Chair: Wendy Reid

Scribe(s): Franco Alvarado

Content:


Wendy Reid: https://www.w3.org/publishing/groups/publ-wg/Meetings/Minutes/2020/2020-01-27-pwg

Wendy Reid: minutes approved

Resolution #1: last minutes approved

1. LPF testing

Wendy Reid: we’ve asked laurent to update on LPF and testing

Laurent Le Meur: here’s a link to a working doc related to LPF tests https://docs.google.com/document/d/1qZ4AMePn6GkLIH2u1LbY50wXgHfGVrCdXhNUXKbDYAM/

Laurent Le Meur: I’ve created a working doc that I’ve formatted in JSON so it’s easier to discuss
… description is an extract from spec with a selection text. the actions are “what should happen to tested package?”
… there will be three or four packages for each test
… errors will be what should happen with this test file or test
… i think ive gone through all of the specification. i have added some tests that should trigger an error
… so most of the tests have errors done. i will make samples that are good

Ivan Herman: +1 for the inclusion of negative tests

Laurent Le Meur: i think that we shall have two tests that cause an error
… i think that ivan you are the most advanced. do you think this is good enough?

Ivan Herman: i can’t judge for the content, i don’t know the document well enough, but this is what i would have done

Laurent Le Meur: so if there is no comment, then i will format in json and make pull request. but you have time to read it and make comments if you like

Ric Wright: LGTM

Wendy Reid: questions or comments on the test now?

Ivan Herman: laurent, you will put it as a top level separate folder?

Laurent Le Meur: yes

Ivan Herman: i will try to duplicate the reporting mechanism that i have for the others to make a separate repo (?) for that as well

2. Implementations

2.1. Readium

Wendy Reid: is there an update on the readium implementation?

Laurent Le Meur: edr lab is an association so the developments are driven by its members and by the grants we get
… because of hte directions given by the members and grants, we are moving in many paths
… one is architectural with readium, the other with thorium (which includes audiobooks), but otherwise we have accessibility focus
… as of today, on readium mobile for read, we have the support for audiobooks with toc. using internal format we use readium
… which is readium webpub manifest. in the code base we can receive an internal format (json)
… similar to webpub format but not exactly the same
… at the moment we can receive json format and read it as an audiobook
… we already have transformation to webpub but its not clean enough to be published yet. not on github yet
… planned for coming months. we will do the same for ios and desktop in the future
… month of june. we would like to present it at the international publishing summit

Wendy Reid: audiobooks do work in the android version. would it be possible run tests against android implementation?”

Laurent Le Meur: at the moment, it would be with the internal format, not the webpub manifest. it would not be adequate

Laurent Le Meur: i would wait for the transformation to be done before testing

Dave Cramer: just to clarify, this transformation from the spec to internal readium format, would that happen inside readium? so that a user could load a standard audiobook (per our spec)?

Laurent Le Meur: yes, but only exception would be when the book is LCP because the format does not support DRM
… the publisher can export the w3c audiobook in here and thats it

Ivan Herman: that sounds great
… we have to have at least six weeks if not two months to be in the clear between the publication of the rec. for that we have to have our testing done
… having the test suite run and reported back does not mean that at that point we are ready to go public with full application
… it only means that the tests and the running implementation prove that the spec is consistent
… i’m trying to find the right timing for this because we will need at the minimum two or more implementations
… to show this consistency
… when could that be possible?

Laurent Le Meur: i understand that the manifest processing test for example, they are testing the spec, not an implementation
… two possible tracks: 1 is to use real life samples of w3c audiobooks made by publishers to check that the spec is right
… using these samples in the test suite
… another totally parallel track is to get two reading system implementations of this spec
… my real goal is to have this ready for june because i want to present it at the epub summit

Wendy Reid: june is too late unfortunately
… testing is due march 31. extendable a few weeks but not much more
… we need to have implementations and publisher tests
… what we need is implementations that can ingest the format

Tzviya Siegman: reminder: our group’s charter is up in july. this a real deadline.

Ivan Herman: to be very specific, for the CR what we need is a real life implementation that goes through all the tests that are in the manifest testing repo
… and goes through those with their implementation and do the results which are expected
… so if you have an internal prototype not yet ready for the public which doesn’t even have the full blown user interface
… thats good enough
… at the time of testing and reporting this is where it is. two months later it comes out with a usable user implementation
… its only internal testing in the sense that the only thing we need is a json report that says yes we ran this one and yes this is the result we got
… it is not necessary that by then the whole project will be complete
… it will be very very precise. for example the TOC is not a normative thing. so if we don’t have that, thats on paper good enough
… the same way as for LPF because these are not normative. but it looks of course better if those are done as well.
… there is no real reason to expect a full blown shiny usable and public implementation to be ready by then
… at some point in time you do this internal testing anyway and maybe during this testing you will come up with more tests

Laurent Le Meur: i will check with the devs what we can do at least in one platform before the end of march
… only limitation may be to treat any toc and html toc that can be done. if we manage to have simple toc, could be ok

2.2. another implementation

Wendy Reid: https://www.jellybooks.com/cloud_reader/audiobooks/little_women

Wendy Reid: i am working on an implementation on jelly books. audiobooks inside cloud reader
… bugging them now to see if they would be willing to run tests
… has toc, player
… they are looking for feedback

Laurent Le Meur: this cloud reader is a prelude to what will be readium web, toolkit later. what you see is good start of what will be in readium later
… totally open source

Wendy Reid: there is going to be in two weeks, i’m doing a webinar on the spec so hopefully there will be publisher attention

Tzviya Siegman: link to BISG webinar https://bisg.org/events/EventDetails.aspx?id=1321302&group=

3. Misc

Ivan Herman: Wendy, what happened with user behavior testing?

Wendy Reid: the PR is there
… need to link to index
… but the tests are there


4. Resolutions