Publishing Working Group Telco — Minutes

Date: 2019-12-02

See also the Agenda and the IRC Log

Attendees

Present: Gregorio Pellegrino, Dave Cramer, Wendy Reid, Ivan Herman, Laurent Le Meur, Marisa DeMeglio, Ben Schroeter, Romain Deltour, Benjamin Young, Garth Conboy, Avneesh Singh, Matt Garrish, Brady Duga, Ric Wright, Charles LaPierre, Joshua Pyle, George Kerscher

Regrets: Bill Kasdorf, Teenya Franklin, Tzviya Siegman, Luc Audrain, Lars Wallin, Nellie McKesson

Guests:

Chair: Wendy Reid

Scribe(s): Dave Cramer

Content:

Gregorio Pellegrino: I’m on the high speed train :)

Wendy Reid: let’s get started

Wendy Reid: https://www.w3.org/publishing/groups/publ-wg/Meetings/Minutes/2019/2019-11-18-pwg

1. minutes

Wendy Reid: minutes approved

Resolution #1: last week’s minutes approved

2. LPF is ready for publication as a Note

Laurent Le Meur: everything is ready
… there was a respec issue

Ivan Herman: pub rules checker had some complications
… so the webmaster will get back to me

Wendy Reid: we’ll see publication this week?

Ivan Herman: Thursday

Wendy Reid: we were approved for publishing CR for audiobooks on Thursday

3. API tests

Wendy Reid: https://github.com/iherman/pub_manifest_api_tests/

Wendy Reid: https://iherman.github.io/pub_manifest_api_tests/

Wendy Reid: https://w3c.github.io/pub-manifest/experiments/manifest-to-internal-processor/testsuite.html

Wendy Reid: I’ve posted links to IRC
… mattg and ivan were working on this

Ivan Herman: there’s nothing new
… I talked about the tests a while ago
… I collected them in one place, independent of my impl
… and have a reporting mechanism
… but this only for one tiny part of what we test, the conversion to the internal representation of the manifest. That’s all it covers.
… implementors can send a JSON file that reports their results, and make a PR; the report is automatic.

4. repo for testing

Wendy Reid: this came up in a PR because we have separate repos for audio and pub manifest
… we could have tests in pwg repo, or in new test repo

Ivan Herman: I raised this issue
… I just have one part of the testing
… we need something similar for json schema
… and there will be other tests
… it’s cleaner if it’s in one repo

Benjamin Young: +1 to one “testsuite” repo

Wendy Reid: I’m ok with a test repo

Ivan Herman: if there’s a better answer for where to put the test I’m all ears

Wendy Reid: everyone has no opinion or is ok with this.

Ivan Herman: there is a pending PR for schema based testing
… if we set up a separate repo, or do the wg repo, then benjamin can populate

Benjamin Young: it’s fine wherever

Matt Garrish: I think having a more general thing like publishing test suites
… having it separate reduces noise from notifications
… so just one for all our specs

Wendy Reid: sounds good

Ivan Herman: ok

Proposed resolution: Create a new repo for tests called publishing-tests and store all of the publication manifest, audiobooks, and other tests there (Wendy Reid)

Geoff Jukes: +1

Dave Cramer: +1

Marisa DeMeglio: +1

Matt Garrish: +1

Joshua Pyle: +1

Brady Duga: +1

Romain Deltour: +1

Laurent Le Meur: +1

Ivan Herman: +1

George Kerscher: +

Wendy Reid: +1

Ben Schroeter: +1

Benjamin Young: +1

Resolution #2: Create a new repo for tests called publishing-tests and store all of the publication manifest, audiobooks, and other tests there

Ric Wright: +1

George Kerscher: copyright on these materials and usage… are we going to just have
… make believe content or real content?

Wendy Reid: public domain or made up
… I’ve been looking for public domain content or make it up

Ivan Herman: one practical question: we already have real examples, are they public domain?

Dave Cramer: Everything I have used is from Librivox

Wendy Reid: I’ve been using librivox, too, for audio
… we do have a testing timeline, with the end of CR being march 31
… we also decided that there are only 2 more meetings this year
… i’d like to have our tests firmed up
… and the remaining part is on me
… so starting in january any implementers can start testing, and we can get results

Ivan Herman: I would like to understand what type of test we will have beyond the one I seeded
… which are only on the transformation part
… beyond that, what other types of tests do we want?
… is there any test for table of contents structure and algo
… and there may be other statements in specs
… and I don’t know what the schema testing does
… and we need tests for some other things

Wendy Reid: the other group of tests is for user agent behaviour, which I’m working on
… self-reported, not automated
… covering the UCR stuff
… and we need TOC tests
… but it’s optional
… so we need mandatory and optional tests
… since we don’t have toc tests, does someone want to tackle those?
… I’ll send an email
… the tests have to be self-run, we don’t have an automated way to report data
… can we do that?
… or is an upload format OK?

Ivan Herman: for what I did, i’m not even sure how I would test automatically
… if this and this is missing, then the test says validation error
… we don’t have a precise set of error numbers like http
… it takes a human
… I don’t think the effort is worth it
… but this is only for the tests I’ve worked on
… I don’t know about user agent tests

Wendy Reid: I wouldn’t automate those
… I’ll send a list of things that need to be done, and hope for volunteers
… we’ve finished the agenda. See you next week!

5. Resolutions