See also: IRC log
<azaroth> scribenick: azaroth
<TimCole> PROPOSED RESOLUTION: Minutes of the previous call are approved: https://www.w3.org/2016/10/28-annotation-minutes.html
<ivan> +1
TimCole: Last meeting was October 28th, any issues with the minutes?
+1
<TimCole> +1
<Jacob> +1
<tbdinesh> +1
<bigbluehat> +1
RESOLUTION: Minutes of the previous call are approved: https://www.w3.org/2016/10/28-annotation-minutes.html
TimCole: There's 3 open issues, most recent was ActivityStream issue
<TimCole> scribenick: TimCole
<Jacob> Agree with Rob. Looks to me like we're doing what we're supposed to.
azaroth: Issue #357, doesn't seem to be an
issue here.
... Activity Streams defines appropriate classes
... where we recommend FOAF semantics, the only thing were implying is
that they overlap with Activity Streams semantics in a couple of areas
... seems like this is okay in RDF, don't see the problem
ivan: where does the concern come from? where do we talk about FOAF?
<bigbluehat> https://www.w3.org/TR/annotation-vocab/#foaf-organization
<bigbluehat> https://www.w3.org/TR/annotation-vocab/#foaf-person
azaroth: in the Vocab we mention
<azaroth> https://www.w3.org/TR/annotation-vocab/#foaf-organization
<bigbluehat> used here https://www.w3.org/TR/annotation-model/#h-agents
ivan: would prefer us to use VCard vocab for some of our keys
azaroth: on the grounds that Activity Streams mentions VCard
ivan: in the mention of FOAF normative?
<azaroth> https://www.w3.org/TR/annotation-model/#agents
azaroth: not a MUST
ivan: is still a borderline because he would ask us to change a SHOULD from FOAF to VCard
<ShaneM> feels like an unreasonable risk
<bigbluehat> +1
<azaroth> +1
<Jacob> +1
azaroth: would end up with mix of VCard and FOAF
ivan: doesn't seem a compelling technical reason to change - tag as postpone?
<Jacob> that sounds reasonable
ivan: too late to change since not essential technically
<azaroth> PROPOSAL: Given no technical reason to change from FOAF to vCard, we will postpone the discussion until a future version
<azaroth> +1
<Jacob> +1
+1
<ivan> +1
<takeshi> +1
<tbdinesh> +1
<ShaneM> +1
RESOLUTION: Given no technical reason to change from FOAF to vCard, we will postpone the discussion until a future version
<bigbluehat> +1
<scribe> scribenick: azaroth
TimCole: Other opens are SVG (raised by Ivan) and registering the profile
Ivan: Can close the SVG issue if we run out of time.
bigbluehat: The only thing so far is changing the reference url to protocol
<bigbluehat> https://github.com/w3c/web-annotation/issues/324#issuecomment-258262342
bigbluehat: This is the proposed submission
... Ivan and Rob said to point to the protocol
... Would be good to add a note to mention the media type in the model
Ivan: Marked as a milestone for rec, as
there's no change to the documents
... It's not even urgent
bigbluehat: With this info, I can send in the submission and report back
ivan: So for PR we have only two open issues.
<TimCole> scribenick: TimCole
<azaroth> https://github.com/w3c/web-annotation/issues/346
azaroth: There is a waiting an addition pointing from model to vocab about media type
<scribe> scribenick: azaroth
https://github.com/w3c/web-annotation/issues/376
<TimCole> scribenick: TimCole
azaroth: 376 has been closed via email
ivan: at risk features will still need to be moved to appendix
<azaroth> scribenick: azaroth
ivan: for the SVG issue can you do it
azaroth: Can do, but the SVG representation generated by Omnigraffle isn't very good
ivan: Okay, I'm happy to just close the issue
TimCole: Features that we need extra work
on. There are 4 criteria we don't have enough implementations of
... Choice of Body, Choice of Target, Creator for Body, Creator for
Target
... For my part, we have an implementation for creator for Body and
Target
... Don't know where we are for a second implementation
bigbluehat: I have the research done and
started coding. The plan is to use a translate API, and let the user
store the body in different languages.
... so a choice for body, and creator (user vs api) for body
<tbdinesh> If its not too late for impls we have list/choice coming soon
tbdinesh: We are almost ready. Maybe a few days. We will have choice and list done.
ivan: When can we get it into the test suite and hence the report?
tbdinesh: I hope by middle of next week
ivan: Benjamin, what about you?
bigbluehat: Mid week would be a more
sensible target than monday
... It's not conceptually hard, but the devil is in the details
ivan: If Dinesh and Benjamin get them into the report, are we done?
azaroth: I'll have a choice and creator of body available, hopefully by the end of hte day
TimCole: So confident for Choice and Creator for body, but not for target
tbdinesh: The obvious implementation is for body
bigbluehat: It's the problem of knowing who
the creator of the target is in the first place
... it wouldn't be hard to get it from a wordpress template, for example
TimCole: I don't think we have the time to do that at this stage though
ivan: This leaves us with features for which we have to declare that the CR exit criteria are not met
<bigbluehat> can we list those specific features (again...)?
ivan: so the only thing we can do is to reissue a CR on the model and the vocab, and put the unimplemented features in an appendix
<Zakim> ShaneM, you wanted to talk about testing todayu and to ask about appendices
TimCole: If we're able to show choice and creator on the body, is there anything in the vocab for it?
azaroth: Nope, as we'll have implementations for the classes
ShaneM: If something's marked at risk, and
we have to remove it, then we don't need a new CR?
... so we could just mark things at risk in a new CR
... and then if there is an implementation, we keep it and if not, we
move it to an appendix
ivan: Yes, but we don't expect any implementations at this point
TimCole: We do the tests before the
criteria, so I don't think we need to change the tests
... They just don't contribute to the criteria
ShaneM: But it'll just clutter the test results with a bunch of fails unnecessarily
<ivan> https://www.w3.org/2015/Process-20150901/#revised-cr
ivan: So we have some time pressure now.
Process wise this ^^ is where we are
... that means that the minimum time is 4 weeks after the publication
... which means we have to be quick in this. We could issue a PR mid
december. Which goes to the end of January. Charter is extended to end
of February
<bigbluehat> so by what date do we need what things implemented?
ivan: So we can do it, but I would like for it to be done quickly
<ShaneM> http://td.spec-ops.io/test-results/annotation-model/less-than-2.html
ivan: We need to vote today that we're doing
this ... that we issue a revised CR
... Then the problem is that I am around next week M/T, not really
around tomorrow, and may need help getting the proper documentation to
the director for the right authorization for a new CR
... so the week after, the week of the 21st, it could be published
<TimCole> PROPOSED RESOLUTION: Reissue CR with Choice for Targets and Agent for Body moved to appendix (or marked as at-risk)
<bigbluehat> so...this only happens if tbdinesh and I don't get our stuff done by Wednesday?
<ivan> PROPOSED RESOLUTION: Reissue CR with Choice for Targets and Agent for Body moved to an informative appendix
+1
<ivan> +1
<TimCole> PROPOSED RESOLUTION: Reissue CR with Choice for Targets and Agent for Targets moved to an informative appendix
+1
<ivan> +1
<Jacob> +1
<TimCole> +1
<takeshi> +1
<bigbluehat> +1
<ShaneM> +1
<tbdinesh> +1
RESOLUTION: Reissue CR with Choice for Targets and Agent for Targets moved to an informative appendix
ivan: So the whole document needs to be
edited for the new status, which is where timing comes in
... do you think you can do this next week, if I'm not around?
azaroth: I can work on the text today, but need someone to look over the process parts of it
ivan: You can look at the CR request from
July
... A publication date of the 24th is a safe bet
... Need to get a reply from Ralph to issue an official request, the
choice is 22nd or 24th
TimCole: 24th is a holiday in the US
ivan: You are right, okay the 22nd
... so the request needs to be out on Monday or Tuesday
... If the whole thing is done, I can send a preliminary publication
request to the web master
... last time I did run into a number of issues with pubrules and so
forth, so please be careful of that
... So a deadline of the 20th of December
... and then there's Christmas, so the PR is January ... PR is 28 days
after publication
... I wonder why the process document uses days not weeks ...
... so PR first week of January, and then in Rec in February
TimCole: CR goes to 20th Dec, but we'll want all of the paperwork done in advance of that
ivan: Mostly the test report
... in the repo there's just the pattern for the emails
<tbdinesh> if anyone can implement choice for target and agent for target, we can do one by end of next week
ivan: in the text there's a description of
the testing process and so on. We should send the request in the week of
Dec 20, which will sit until the first week of January
... lets get the CR done first, that's the urgent thing
TimCole: So the good news is we dont' need the test reports next week
<Zakim> ShaneM, you wanted to talk about testing today (so I dont forget)
ShaneM: Someone mentioned tests today.
There's a problem with the tests in the main repo, so use mine
... We removed the lines from the tests that said ignore the results of
optional tests. So the test reports are currently wrong
... I made the change back and re-re-ran the tests, but need to get them
merged to the main repo
<ShaneM> http://td.spec-ops.io:8000/tools/runner/index.html
ShaneM: That server reflects the fixes as
above, and the changes for collection and pages that Tim made yesterday
... The report currently looks like ...
<ShaneM> http://td.spec-ops.io/test-results/annotation-model/all.html
ShaneM: this is a report that is rolled up.
There's a column per implementation.
... the tests are alpha sorted, which puts them out of order
... the optionals in yellow in the second part. The list of tests are
essentially in a random order in the results
... so we re-order in the report
ivan: Red vs Yellow?
ShaneM: Red is that it actually failed
... Not entirely sure what it means in practice
TimCole: Not sure why my and Takeshi
implementation got fails not yellow
... The tests aren't very sophisticated, so I'm not sure how it would do
that.
... it has to have a type, and from an enumeration
... if you don't have the key at all, it should report that you failed
... we can investigate this later
Ivan: Looking at the protocol tests
... it fails on two required things
azaroth: I can fix those
<ShaneM> what's the URI to your implementation Rob?
Ivan: We have several weeks because of the
CR issue
... what about vocab?
ShaneM: Unfortunately, I don't have time to put the report together, but the results are there so can someone else create a report?
TimCole: I'm happy to help put it together... but I don't know what was really done and how to make the report
ShaneM: If it's more than a few paragraphs I think we're doing it wrong
<bigbluehat> this is where it goes though, yeah? https://github.com/w3c/test-results/tree/gh-pages/annotation-vocab
ShaneM: That's where it goes
azaroth: I'll dig up my code
ivan: The CR republication is the urgent thing
bigbluehat: Let's make a task list and when they're due by
ShaneM: Hugo has been running the protocol tests, and we should have results from them
ivan: That would be three which is good
TimCole: Collection and Pages... anyone
satisfying the protocol tests should be able to submit results
... I don't see them in Shane's new test server
ShaneM: We don't have any results
... I don't have input for them
TimCole: Rob you can submit one?
ShaneM: The other is the test suite itself
... is that okay?
ivan: Yes that's fine
ShaneM: We can ask Hugo to run theirs
TimCole: Where we are with the spreadsheet
from earlier this week
... the column with the identifiers might need updating
... to map which tests map to which criteria
... should discuss to make sure I got it right, at the meeting after
thanksgiving
ivan: Once we've verified, put it into HTML
or markdown and put it up along side the test results
... and link to the table
TimCole: Sounds like a plan
... one last thing, there's some errors with the Pund.it implementation
in terms of context
... which is required for the model. Where do we sit on that?
<ShaneM> its okay that they fail that test...
azaroth: I can talk to them, but we have the right number of implementations
ivan: Yes, that's their problem, not ours in
some sense
... Not about testing, but to be sure that we know what the next steps
are
... Rob to take the lead on sending the request for new CR to Ralph and
Phillipe
... There has to be 5 business days between request and publication
... There must be a publication moratorium list somewhere
TimCole: Would monday be in time for 22nd?
azaroth: Okay, so I'll do the editing today and over the weekend and the request on Monday
ShaneM: Could be a bug in the report generator for the fails
<ivan> https://lists.w3.org/Archives/Member/chairs/2016AprJun/0004.html says there is no moratorium on thxgiving
<ivan> azaroth, TimCole, see ^