Web Annotation Working Group Teleconference

<ivan> scribenick ben_thatmustbeme

<bjdmeest> scribenick bjdmeest

<ivan> scribenick: bjdmeest

<TimCole> PROPOSED RESOLUTION: Minutes of the previous call are approved: https://www.w3.org/2016/11/11-annotation-minutes.html

TimCole: [talking about agenda]
... any objections about the minutes? no?

RESOLUTION: Minutes of the previous call are approved: https://www.w3.org/2016/11/11-annotation-minutes.html

issue updates

TimCole: 2 open issues

<TimCole> https://github.com/w3c/web-annotation/issues/381

TimCole: ... #381 is the first one
... previousVersionURI didn't exist in OWL

ivan: I did handle that
... PR waiting for approval
... #385
... replace non-existant with ??? from prov, which came closest imo
... @azaroth, could you have a look? there were a lot of changes

azaroth: looks good to me. closedQ!

<TimCole> https://github.com/w3c/web-annotation/issues/386

TimCole: any other comments about that? no? moving on...
... #386

azaroth: testing question
... there is a company in the UK that has implemented the model and protocol
... they implemented the IF-MATCH check
... but there's a problem with the test
... the ETAG header of the response should match

ShaneM: is this #386?
... I missed that, it looks like a fine thing to do, I'll do this today

ivan: can you close after you did that?

ShaneM: no, don't have write access, I'll ping one of you

TimCole: is there any doublecheck that Shane can do?

azaroth: my server had it implemented but switched off, I'll turn it on once it's implemented

ShaneM: I'll let you know, so you can test yourself

azaroth: I have the report of that company from the UK
... I could wait and let them retest once ETAG is fixed
... change the JSON manually

ivan: I'm in favor of the last one

CR status update

TimCole: we proposed a new CR on the vocabulary and the model on the 22nd of November

azaroth: changes:
... we took out the AT RISK features of composite, list, and independents
... because of insufficient implementations
... only one of list and composite

azaroth: [going over the changes]

TimCole: about the context document?

azaroth: we can just change that one
... we took independent, composite and list out of the context
... people that are using them, those terms will no longer correctly resolve into RDF

TimCole: any comments about this?

<bigbluehat> could we publish them as a separate context? for folks who want them?

<bigbluehat> or is that too hard...?

azaroth: this needs to be turned into an extension for it to keep working

ivan: the only thing we may want to do is record an issue that says that it would be good to have that in V2

<tbdinesh> bigblue.. like an evolving context

ivan: we may come back to that later

<tbdinesh> bigblue.. like an evolving context standard

TimCole: we will remove from the Model tests any tests that use composite, list, or independent
... or those that mention choice on target
... I'll get that done, probably over the weekend
... anything else about CR?
... we have a lot of promising test results
... for all other features

Need to have PR Transition ready by 20 December

<ivan> https://github.com/w3c/web-annotation/blob/gh-pages/admin/PRTransitionAdmin/PRTransitionRequest.md

ivan: I made a draft of the transition request
... what we need to do (except for having the finalized 3 reports)
... really important are 2 things
... 1: features marked at risk

<bigbluehat> Previous topic (sorry): here's the link to resurrect Composite, Independents, and List: https://github.com/w3c/web-annotation/issues/387

ivan: maybe @azaroth can refine the text to reflect what exactly happened
... and 2: specify what happened with the activity streams

azaroth: it has CR 6/12

ivan: anything new?

<bigbluehat> AS2 remaining issues https://github.com/w3c/activitystreams/issues

ivan: we have to contact them
... we can refer to CR as a PR, but by the time we have a recommendation, their spec also needs to recommendation
... so we need to be sure

azaroth: I'll ask around, at least James

ivan: what's our fallback if we have problem?

azaroth: we can just duplicate all classes and predicates into our own namespace

ivan: as we said? only changes to the vocabulary document wrt the recommendations?

azaroth: yes, we don't mention AS in the model, so that's fine

ivan: can we set a deadline for this?
... let's try to get that sorted out in a week?
... if not, then we have to change the documents as well
... I would propose: if we don't have clear information about the AS2 PR transition by, e.g., the 16th, we take the fallback measure

<ivan> Proposal: if we don't have clear information about the AS2 PR transition by, e.g., the 16th, we take the fallback measure

<azaroth> +1

<ivan> +1

+1

<Jacob> +1

<TimCole> +1

<bigbluehat> +1

<tbdinesh> +1

<ShaneM> +1

<ShaneM> I doubt that it will come to a shock to W3M that coordination is difficult

RESOLUTION: if we don't have clear information about the AS2 PR transition by, e.g., the 16th, we take the fallback measure

ivan: next: all implementation information section should be as complete as possible
... relying on the fact that all features have at least 2 implementations
... there is a whole line where only implementation passes

azaroth: there are test results pending that implement that

<bigbluehat> discussing these results v

<bigbluehat> https://w3c.github.io/test-results/annotation-model/all.html

azaroth: also for the other

<bigbluehat> https://w3c.github.io/test-results/annotation-protocol/all.html

ivan: [so, all features will have at least 2 implementations]
... [about the test results]: some descriptions are a bit too cryptic

TimCole: there are 2 approaches
... there are assertions and test patterns
... the assertions cover a lot more than what we need to exit CR

<ShaneM> I am confused. Are we talking about model or protocol?

<bjdmeest> scribenick: bjdmeest

ivan: I like the approach of creating a separate html page, specifically for the transition
... we have to be careful to send out the correct url

ShaneM: didn't we send the README.MD as url?

azaroth: [no]

ivan: whatever is quicker and easier... we just need to have it done

TimCole: we now have an ID on all subsets, and spreadsheet with all results
... it is easy to create a new HTML
... to show only what matters for the CR
... that's easy to create

ivan: and an important one

<Zakim> ShaneM, you wanted to talk about an approach

ShaneM: the top of that report is arbitrary html, so we could add the rolled up implementation report there
... or a link to the external rolled up html

TimCole: how to add the rolled up to the report?

ShaneM: there's a file 'report-header.html', you could just change that
... the magic and unicorns will do the rest

ivan: about the rolled up html: can we complete this by the end of next week?

TimCole: I can try to do this on Monday or Tuesday

ivan: can I dream about a situation, by the end of next week, where all specs are perfect?

TimCole: we need extra test reports, reports of the UK, and @azaroths work

azaroth: I can do it whenever

TimCole: next week is a reasonable target

ShaneM: I wrote an automated tool to rerun all the manuel tests, and regenerate all the results
... once we have the tests submitted, I can press 'go'

azaroth: I think we need to distinguish between collections and annotations
... in the collection MUSTS, that fails, because they test collections, not annotations
... we shouldn't test annotations and say they pass collection tests

TimCole: it would be nice to distinguish that RI and the next two are implementations of collections and pages

<azaroth> In essence, we are saying we pass a bunch of tests (the collection and pages musts and optionals) because we're running Annotations through those tests,and they have similar properties, or the tests check for presence first and (clearly) total is not present on an Annotation

<azaroth> So we pass tests not through actual implementation but by testing data that isn't intended for that purpose but happens to pass the tests, as written

ivan: so, can we talk about dates for getting the deadlines? or are there other things?

TimCole: there are some tests, some rewriting... it can be done in a week

ivan: in 2 weeks, on the 16th, everything is closed
... by closed, I mean, we have no pending issues about the transition
... first of all, there is christmas time coming, I will be on vacation starting the 20th, the week of the 9th is complicated... It will be difficult to find the right transition date
... also, the PR should be up for 4 weeks, and our charter ends on february
... so before the 20th, I send the transition request for PR
... and we try to find a transition date by the end of January
... so we have a transition around the 5th or 6th
... the goal would be to publish PR the 17th of January, or 24th is still ok
... so, I'd like to have everything ready in two weeks, to have the process started asap

<ShaneM> =1

TimCole: who is available the 16th?

ivan: we have to have a formal vote, we cannot have this vote before the 20th

TimCole: we'll do as much as possible by next week

ivan: the PR versions will only need to be done by January, ReSpec will do most of the work

TimCole: we needed to talk about the transition requests, whether there'll be a new WG, or the open annotation WG

<bigbluehat> +1 to reusing things

ivan: my preference would be to reopen the Open Annotation CG
... I need availability dates for you guys for the transition request
... 5, 6, 10, 16

TimCole: fine for me

azaroth: also fine

bigbluehat: also fine

ivan: I'll put in right now an entry to Philip's and Ralph's calendar, so we can act in between

TimCole: so, one pull request pending, I will the make the deletions according to that pull request
... next week, no full group meeting, only if test relevant stuff

<ivan> trackbot, end telcon