Web Annotation Working Group Teleconference

<azaroth> trackbot, start meeting

<azaroth> PROPOSED RESOLUTION: Minutes of previous call are approved http://www.w3.org/2016/01/06-annotation-minutes.html

<fjh> s;Agenda.*;Agenda: https://lists.w3.org/Archives/Public/public-annotation/2016Jan/0036.html;

<fjh> i;Agenda;Agenda: https://lists.w3.org/Archives/Public/public-annotation/2016Jan/0036.html

<fjh> Agenda: https://lists.w3.org/Archives/Public/public-annotation/2016Jan/0036.html;

<azaroth> Scribenick: bjdmeest

azaroth: agenda: planning 2016, and next 4 issues

shepazu: [talking about iAnnotate in Europe]

azaroth

azaroth: Social Web released some FPWD's
... most interesting one is WebMention

<azaroth> http://www.w3.org/TR/webmention

azaroth: it's a very simple notification network
... to send notifications between systems that annotations have been created or modified
... which is one of our use cases

shepazu: it builds upon existing things and tools
... advantageous for us to reuse

<azaroth> PROPOSED RESOLUTION: Minutes of previous call are approved http://www.w3.org/2016/01/06-annotation-minutes.html

<shepazu> Also, Social Web WG published Social Web Protocols https://www.w3.org/TR/2016/WD-social-web-protocols-20160112/

RESOLUTION: Minutes of previous call are approved http://www.w3.org/2016/01/06-annotation-minutes.html

Planning for 2016

azaroth: discussion started last week
... about planning
... oustanding issues:
... 1. f2f timing, amount of f2f's, and their purposes
... we need to be clear on what deliverables we will be working on, to what end state
... proposed model doc to go to CR
... we should timebox, so we can make progress

shepazu: [talking about F2F] in discussions with Dan, we talked about a 1 day F2F
... F2F are valuable and expensive, but I think we should have at least a 2 day F2F

fjh: makes sense
... Implementations and test cases for Model should be lined up, to go to CR
... should be important for F2F, depending on how many we can add to the room

shepazu: fact that we are co-allocated with iAnnotate, could help to open the room for half a day
... could be useful for iAnnotate participants, and for us
... e.g., last day of F2F
... dedicated to an open session with the community

azaroth: timing of iAnnotate would be May

<fjh> we need to be clear that we'd like to get model to REC by end of charter, others possibly entering CR (or in old terms in LC)

<davis_salisbury> sorsorry, something is wrong with the client

<fjh> s/sorsorry.*//

shepazu: testing is superimportant, and takes a lot of time
... we need to find some specific topics to be addressed in person
... let's avoid too much testing at F2F

chrisbirk: [about testing plan]
... maybe bring it to the group next week?

shepazu: we'll be facing a similar problem as WebPayments WG
... am talking with that group's testing lead
... maybe we can share infrastructure

azaroth: so a plan by next week?

chrisbirk: yes

ivan: one thing we have to do on the F2F: charter expires end of september
... continuing without problems would involve thinking about rechartering by F2F
... F2F would be a perfect place for that

azaroth: can we make an agenda wiki page? logistics, guests, etc..

ivan: doug's point is valid, that we have to plan for 2 days
... will be in Berlin, probably before iAnnotate (Tuesday, Wednesday)

azaroth: what should we work on up until then?
... 1. Model: split up in vocabulary as well
... 2. protocol, and 3. findText
... Model and voc can go to CR, what about the others?
... we need tests, testing framework and at least two independent implementations

fjh: findtext seems really important, and gets buy-in of implementers
... what can we do, even without going to CR
... about protocol: what happens if we don;t get it done?

shepazu: there is interest in having something like findtext
... we can leverage that
... second point: how?
... form a process and chartering perspective: if browser vendors say 'yes we're interested' would be sufficient for us
... as long as we keep iterating and improving the spec
... maybe we can start some testing

<fjh> makes sense

shepazu: with browser vendors interest, this will be deemed viable work

<Zakim> azaroth, you wanted to +1 protocol

azaroth: protocol +1
... in order to have interoperability, we need to be able to transfer annotations between systems
... e.g., maintainer vs rendering system
... only data model would not suffice
... I think protocol is very easy to implement
... basically it is just REST
... also testing would be straightforward

<fjh> +1 to be able to transfer annotations

azaroth: documentation will be close to done after implementation and testing

timcole: I agree protocol being important
... I am a little worried about (relates to findtext): we don't have a clear understanding about how an annotation will look like in html
... don't know whether protocol will address that
... do we need something explicit about html-centric rendering of annotations?

azaroth: it's been on our books for a while

ivan: speaking of protocol --> speaking of functionality level of protocol document, not about search?

azaroth: indeed, get update delete

ivan: in the coming months, without clear idea about search, we cannot have a recommendation about search by september

fjh: concerned about complexity: LDP, paging, synchronizing...
... there are a lot of discussions that could take a lot of time

shepazu: how is the protocol is different than LDP, concerning annotations?

<fjh> need to scope work carefully to have CR deadline this year - noting goals related to synchronization, containers, paging, searching, profiling other work etc

shepazu: also: I did not expect an annotation specific protocol, rather reusing an existing one, e.g., from social web wg

<bigbluehat> how close are we to LDP? are we that different outside of some "say these few other things"?

<bigbluehat> JSON-LD by default (not Turtle)...and some others :)

azaroth: we changed making turtle not required
... we are more opinionated about the response than LDP
... and more opinionated about what can be in a single container
... these are profiles of LDP
... not something new

shepazu: it's advantageous to us if we can reuse existing implementations
... have we exhausted all options?

azaroth: if search is so important, then we don't have any options

<bigbluehat> you can use LDP "out of the box" for annotations, correct? we're just trying to lower the bar on implementation--via the default changes and narrowing the scope to annotations?

<fjh> syncronization across servers?

azaroth: issue around paging: pages in LDP are not ordered internally, you have to reorder within single page
... there are several use cases where that is not possible
... previous decision was using ordered collection from activity streams

shepazu: is there anything else we can reuse?

azaroth: correct
... that we cannot
... can we agree on a resolution on what we can take to CR?

shepazu: we are a long way atm

<fjh> +1 to doug, need to work out details of protocol scope, details etc

<tilgovi> +1 to doug

ivan: we have a plan, an idea about what can go, and we'll work along those lines

<azaroth> +1

<fjh> +1 to Ivan

<shepazu> Wiki page for F2F: https://www.w3.org/annotation/wiki/Meetings/F2F_Berlin_2016

issue 49

<ivan> https://github.com/w3c/web-annotation/issues/49

https://github.com/w3c/web-annotation/issues/49

azaroth: issue is about how systems can notify each about existence and modifications of annotations
... proposal: postpone, given WebMention's recent FPWD
... we need to get reading and writing done before we can think about notification

ivan: choice is whether we do something about it now, or label it as issue that we tackle it in new charter

<shepazu> https://www.w3.org/annotation/charter/

shepazu: it's not on our charter, so shouldn't be taking on work
... see no argument to do it
... hopefully, we can rely on the work of webmention
... although webmention might not fully comply to our needs
... there are other kinds of notifications as well

<azaroth> PROPOSED RESOLUTION: Defer technical discussion and work on notification, continue to work with Social Web WG to ensure requirements are met

<Jacob> +1

<ivan> +1

<azaroth> +1

<davis_salisbury> +1

<tilgovi> +1

<tbdinesh> +1

<tilgovi> de-prioritize?

<tbdinesh> +1 to doug

<azaroth> PROPOSED RESOLUTION: Defer technical work on notification, continue to discuss requirements and work with Social Web WG to ensure those are met

<azaroth> +1

<shepazu> +1

<davis_salisbury> +1

<chrisbirk> +1

<tbdinesh> +1

<ivan> +1

<tilgovi> +1

<TimCole> +1

RESOLUTION: Defer technical work on notification, continue to discuss requirements and work with Social Web WG to ensure those are met

<Jacob> +1

issue 59

<ivan> https://github.com/w3c/web-annotation/issues/59

https://github.com/w3c/web-annotation/issues/59

azaroth: in model, we talk about normalized text (for the offset selector)
... this issue is about how this works in the DOM
... should that be a normative reference?
... I have no objection to that
... any objections?

<tilgovi> this is a PR for the spec language change: https://github.com/w3c/web-annotation/commit/bed9d0aeace6c27c32e7c8b6f04a7afc8766d5b7

<ivan> works for me...

<tilgovi> or, rather, that's a commit that could be merged

azaroth: we can merge that

shepazu: about HTML serialization: can we put that on the agenda for next week?

azaroth: yes

issue 21

<ivan> https://github.com/w3c/web-annotation/issues/21

azaroth: about set of uris that would be listed in `via`: ordered or a set?
... has implications for model
... ivan suggested list makes sense (i.e., breadcrumbs)
... makes it harder to use viana:via
... as that is just a regular predicate, no list

<TimCole> slight preference for ordered list

ivan: I would say: go for ordered, if major issues, we deal with it

<ivan> rrasagent, pointer?

azaroth: agreed
... anyone against ordered list?
... so: we go for ordered, and if feedback, we can set it back
... [wrapping up]
... for next week: HTML serialization and testing

<ivan> trackbot, end telcon