Web Annotation Working Group Teleconference

<fjh> trackbot, start telecon

<trackbot> Date: 11 November 2015

<azaroth> Chairs: Rob_Sanderson, Frederick_Hirsch

<fjh> Agenda: https://lists.w3.org/Archives/Public/public-annotation/2015Nov/0178.html

Agenda Review, Scribe Selection, Announcements

<azaroth> Thanks Tim!

<fjh> ScribeNick: TimCole

Minutes

<azaroth> proposed RESOLUTION: Minutes from 4 Nov approved: http://www.w3.org/2015/11/04-annotation-minutes.html

RESOLUTION: Minutes from 4 Nov approved: http://www.w3.org/2015/11/04-annotation-minutes.html

azaroth: Any announcements?

shepazu: following up from TPAC with those who might be interested in the FindText API or other annotation issues
... showing them mock-ups of what annotations might look like in their browsers.

Working Mode

<azaroth> Link: https://lists.w3.org/Archives/Public/public-annotation/2015Nov/0088.html

azaroth: linked doc outlines methods for making faster progress (as compared to this summer)
... try to take issues more breadth rather than too much depth
... so we can move on issues that we can get consensus, rather than getting stuck on smaller issues that on which we can't get consensus
... will support more incremental decisions
... so we will try to stay to 15 minutes per issue
... if consensus, great, if not keep on stack for revisiting in the future.

shepazu: 15 minutes to discuss the issue on the call, assume that we will have conversations on github and the list before issue comes to call for 15 minutes
... is this what you are proposing? Or do issues go immediately to call once raised on github

azaroth: the former - we should continue to have conversations outside the calls and then quickly finalize consensus on the call when ready
... as much consensus as possible off-call before bringing to the call

shepazu: since W3C is doing more asynchronously, we should be sure to call out each decision on list or github rather than relying solely on minutes.

azaroth: any other comments for or against?

no one raised their hands

azaroth: So, hearing no further comments, we should try this new and improved practice.

<azaroth> proposed RESOLUTION: We will endeavor to confirm consensus around github issues in at most 15 minute time-boxed discussions on calls, per https://lists.w3.org/Archives/Public/public-annotation/2015Nov/0088.html

<azaroth> +1

<shepazu> +1

+1

<takeshi> +1

<bigbluehat> +1

<fjh> +1 with chairs judgement as needed

<PaoloCiccarese> +

<PaoloCiccarese> +1

RESOLUTION: We will endeavor to confirm consensus around github issues in at most 15 minute time-boxed discussions on calls, per https://lists.w3.org/Archives/Public/public-annotation/2015Nov/0088.html with chairs reserving right to adjust as appropriate in real time

Splitting annotation model into 2 parts

<azaroth> Link: https://github.com/w3c/web-annotation/issues/97

azaroth: Web Annotation and Web-annotation-model, was discussed at TPAC
... issue raised was that specs aren't satisfying to people interested only in JSON implementation nor to those focused on RDF/linked data and ontologies
... so we looked at approach being used by Social WG re Activity Streams
... they also want to make sure their outputs are useful to both communities
... another advantage (suggested by Ivan) currently we have one doc that includes Turtle and JSON-LD, for testing would be helpful for testing to separate

fjh: confused about 2 things
... would model and serialization still be in one document or would they be separate?
... which terms would appear only in one, which in both?

<tilgovi> +0

azaroth: Web annotation is just about JSON serialization and would have to include structure
... second document would describe the ontoogy -- all of our terms and terms from other vocabs on which we rely

fjh: the model doc would be the RDF and not have JSON?

azaroth: basically yes

<fjh> ok, so we would have a stand-alone JSON-LD document + a model document similar to what we have now (with some JSON-LD removed)

shepazu: I think I am in favor of this and this addresses a long-standing concern, but need to see how this will look before confirming that I'm in favor of it

<bigbluehat> https://github.com/w3c/web-annotation/pull/99 has progress toward this document split

shepazu: need to see to make sure something doesn't get left out, e.g., don't want to have to look in both docs to know that you are a conforming [JSON] implementation
... want to make sure that the conformance issues are all addressed in one document
... 2nd would like to revisit the naming of the documents, Web annotation doesn't seem specific enough

<fjh> why not "Web Annotation Model" and "JSON-LD Web Annotation Serialization"

shepazu: Web annotation might sound too broad. Need to communicate easily.

azaorth: Annotation Model and Annotation Vocabulary?

shepazu: seems in the right direction

<fjh> is vocabulary code for "RDF stuff"?

<azaroth> fjh: yep :)

paoloCiccarese: kind of like model and vocabulary, thought model will include some vocabulary
... the connection between the two docs is going to have be the @context
... how do you talk about the @context in model when the terms in the @context are in the vocabulary?

shepazu: Web annotation model need to include terms we expect to see in the JSON, we can still mention the @context in the Model and say see Vocabulary spec for more details about the terminology

<tilgovi> If the model still needs to define the space, I don't think it can really exist without vocabulary terms in it. They'll have to be words about targets and bodies and susch. But I guess it doesn't need tables of definitions, CamelCaseClassNames and namespaces?

shepazu: so we have some normative vocabulary in the Model document, but make it concrete and complete in the Vocabulary spec.
... would this approach work?
... we might need to look at the specs to figure this out

azaroth: Propose that the editors make progress on splitting and then bring back for further discussion and refinement

paolo: We could just not say much or anything at all about @context in the Model, and talk about @context in the Vocabulary

fjh: I think this could get hopelessly confusing
... maybe we need 3 docs: Model, JSON serialization doc, RDF doc to explain all the concerns for RDF audience
... leaving any vocabulary out of the Model seems a problems.

shepazu: I think we are talking about much the same thing, but we have differences in exactly how to split out
... one issue would be how to talk about how we connect with / use outside vocabularies

<fjh> need to see something to understand what it is

shepazu: but let's make a resolution to let the editors propose something concrete

fjh: if the editors can do a couple of sections that might be enough to make concrete

paolo: as a developer, if I look at the model and I see JSON, it wouldn't be a problem not to know why the provenance pattern is the way it is, that's okay with me as a developer

<azaroth> proposed RESOLUTION: Editors to generate sketch of the split specs for further discussion

paolo: I think the editors can have a separate discussion

azaroth: I think we have a consensus

<fjh> ACTION: azaroth ot work with other editors to create sketch of new document split to make it concrete [recorded in http://www.w3.org/2015/11/11-annotation-minutes.html#action01]

<trackbot> Created ACTION-30 - Ot work with other editors to create sketch of new document split to make it concrete [on Robert Sanderson - due 2015-11-18].

TimCole: do we need more editors given the extra documents?

azaroth: I think the 3 current editors could have a sketch by next week, and then see if more help will be needed?

<azaroth> +1

+1

<shepazu> +1

<fjh> +1

<PaoloCiccarese> +1

RESOLUTION: Editors to generate sketch of the split specs for further discussion

Timeframe for Social WG to come to a solid AS spec so as to support what we want to use

azaroth: Idea is to give Social time to come up with something, but we don't want to be locked / blocked. If they can't do this by our need date, we'll create our own micro-ontology for now
... we would need by mid-January to make decision.
... by March we need a stable ontology, one way or another
... mostly effects annotation collections
... other possible overlaps -- embedded content

shepazu: This seems reasonable, and Social WG at TPAC seemed to see this as reasonable
... but has the Annotation WG told the Social WG what our requirements are? Have we written these down (even informally)?
... maybe we should include in the proposed action?

<fjh> +1 to sharing with Social WG the needs and time frame

shepazu: I haven't found an exact definition of what we need from Social.

azaroth: we have discussed, and they've been good about accommodating our needs
... I have talked with James S. about adding a language attribute, but they have a slightly different use, so may not be able to have language in their model in a way we can use
... no conflict, just less aligned
... but yes, what we are missing is an articulation of the exact requirement
... to some degree these are administrative, e.g., do we need them at CR level by March

shepazu: I think more important to clarify the exact technical requirements, can we write this down?

azaroth: We have been talking more on their turf, so yes we should write down and make visible from our side

fjh: Along similar line, did azaroth get a sense of their current timeline

azaroth: not really, Social didn't meet at TPAC, we talked to chairs
... Social WG has a f2f in December in San Fran and azaroth has joined and plans to attend f2f
... this will help get a better sense of their timeline.
... Rob & Benjamin should write up the points of alignment, bring back to Annotation call, then to Social

<fjh> this should help greatly, thanks Rob

shepazu: my sense at TPAC is there is not yet agreement within Social as to whether they will be able to meet timeline

azaroth: my sense from their calls is that they have a strong divide also along the JSON / RDF-linked data lines
... James S. has had some success in finding the points of consensus, and so I am hopeful that AS will get to the point we need, other Social deliverables less clear

Do we have consensus on closing the following?

<azaroth> proposed RESOLUTION: Close https://github.com/w3c/web-annotation/issues/103

<shepazu> +1

<azaroth> +1

azaroth: 1 can we close http vs. https? Seems consensus that this is out of scope for us.

+1

RESOLUTION: Close https://github.com/w3c/web-annotation/issues/103

<PaoloCiccarese> +1

<azaroth> proposed RESOLUTION: Close https://github.com/w3c/web-annotation/issues/98

<PaoloCiccarese> +1

<azaroth> +1

azaroth: 2. Can we close the issue re reasoning with annotations? The specs we do should remain silent about this.

shepazu: does this mean that the vocab/ontology spec doesn't talk about reasoning?

azaroth: yes. vocab/ontology should stay simple rather than trying to nail down all the reasoning issues.

<shepazu> +1

paolo: it's not the annotation body itself but it's the RDF that is carried with the annotation. we are talking about the structure which does not necessarily carry enough for reasoning

shepazu: there could be some work in the future on this, but for now let's close

azaroth: out of scope for the current work

tilgov: Had some difficulty following the conversation on this
... the choice of which resource ends up the target and what ends up as the body unless you have some kind of explicit relationship between the two, possibly in lieu of roles
... but can close for now and revisit later.

RESOLUTION: Close https://github.com/w3c/web-annotation/issues/98

azaroth: we are adjourned.