W3C

Web Annotation Working Group Teleconference

18 Mar 2015

See also: IRC log

Attendees

Present
Frederick_Hirsch, Rob_Sanderson, Ray_Denenberg, T_B_Dinesh, Dave_Cramer, Matt_Haas, Paolo_Ciccarese, Kyrce_Swenson, Ben_De_Meester, Tim_Cole, Davis_Salisbury, Doug_Schepers, raphael, Randall_Leeds_(at_social_web_F2F_on_IRC_only), Kristof_Csillag
Regrets
Benjamin_Young_(at_social_web_F2F), Bill_Kasdorf, Ivan_Herman, Luc_Moreau
Chair
Frederick_Hirsch, Rob_Sanderson
Scribe
dauwhe

Contents


<trackbot> Date: 18 March 2015

<fjh> Agenda: https://lists.w3.org/Archives/Public/public-annotation/2015Mar/0045.html

<scribe> scribenick: dauwhe

Agenda Review, Scribe Selection, Announcements

fjh: we should start
... add to agenda update on social f2f, randall and benjamin were there, too

Minutes Approval

fjh:

<fjh> proposed RESOLUTION: 11 March 2015 minutes are approved,

<fjh> http://www.w3.org/2015/03/11-annotation-minutes.html

RESOLUTION: 11 March 2015 minutes are approved, http://www.w3.org/2015/03/11-annotation-minutes.html

F2F meetings

shepazu: TPAC is in Japan
... Rob will be there

<RayD> no (TPAC)

<fjh> who can attend tpac?

shepazu: some people may not know what TPAC is
... TPAC is an extravaganza of standards!

<azaroth> http://www.w3.org/2015/11/TPAC/

shepazu: w3c's all hands meeting
... most WGs get together
... groups will meet by themselves, and then also have time for joint meetings with other groups
... you might coordinate with groups you might not normally interact with
... it's in a different location every year
... last year was Santa Clara
... this year i

<csillag> csillag has joined #annotatit's inon

shepazu: Sapporo, Japan
... there will be monkeys

<fjh> http://www.w3.org/2015/11/TPAC/

shepazu: geographical balance helps more people attend at least some w3c meetings
... not a lot of Asian participation in this group
... this is a good opportunity to meet with Asian companys/projects interested in annotations
... Kobo ebook reader is based in Japan, Sony also has an ereader
... we should look at Asian market for annotations
... and let people know beforehand that we will be meeting

<fjh> we should do prep work for TPAC re Asian market

<azaroth> +1 to Doug. W3C is a global community :)

fjh: [reminds folks to Present+ Firstname_Lastname]
... it's a global thing
... we can do questionairre on time
... can you set that up, shepazu?
... F2F in April; needs to know who's coming

<shepazu> https://www.w3.org/2002/09/wbs/73180/annotation-q1_2015/results

fjh: might want to deal with social stuff

shepazu: you asked about existing registration
... 15 people will attend, 8 won't

azaroth: I put that on agenda to confirm that poll
... that people had booked travel, etc

shepazu: once we're at the poll stage, we'd already agreed about location and having a F2F
... we're not reopening that

<fjh> +1, April F2F is set

fjh: we need to have an agenda

shepazu: all of us can see the poll

<RayD> link to poll?

shepazu: some said "I'll come if I get funding"
... we'll discuss offline about funding

<RayD> thanks rob

shepazu: let's confirm with David Salisbury and Kyrce and Matt_Haas about funding
... please update poll if your plans have changed

fjh: Let's set up draft agenda offline, share proposed agenda on list

azaroth: just want to raise the point that we need an agenda

TimCole: questionairre did ask about LDP F2F
... is that worth attending?

azaroth: that will be a technical discussion about LDP and its uses
... and how to move forward on related topics
... what will drive further adoption

shepazu: we should get response from meeting planners about meeting at tpac

<azaroth> Technical discussion about LDP features and usage, related technical specs on Paging and Patch, plus new charter for LDP 1.1

shepazu: do we have a chair who could attend

azaroth: I can go if there's enough people coming

shepazu: Ivan and I should not count for that metric

fjh: we have a deadline on that

shepazu: so we have one chair, I'll set up a poll about interest in attending

fjh: more if you think you will attend
... when's the deadline?

shepazu: we have some time

fjh: we need to know sooner rather than later

Social Web

fjh: takeaways from yesterday discussion
... there was open issue about json-ld
... that activity streams 2.0 spec includes json-ld normatively
... not all implementations may want to process
... json-ld is there, context is defined

<fjh> social web WG defines JSON-LD and @context, but does not require @context to be present, can be implicit and ignored by apps that do not care, reasonable

fjh: bottom line is that json-ld is usuable, but it's ok if you don't care

azaroth: this is great news
... there was lots of talk about not doing json-ld at all
... this is really important

fjh: what Tantek was saying is that if you don't care about this stuff it's not a problem
... the 2nd topic is related to the protocol
... narrowing down their approaches
... question whether LDP should be in that list
... the micropub is the leading options
... form-based encoding
... tied into existing implementations
... a community that supports LDP
... looking at up and downsides of micropub
... bottom line: still in play but might not be LDP

<azaroth> pump.io was also on the list of protocols

fjh: so should we extend micropub?
... we made no promises, but we were considering LDP as basis

azaroth: it would be valuable to look in detail at both miscropub and pump.io

<fjh> http://indiewebcamp.com/Micropub

azaroth: what range of annotation options we want to allow

fjh: you can extend vocab by adding h-types
... but might not make sense
... third topic is there's an issue there's some overlap between groups
... some social web activities might be seen as annotations (like +1)
... we need to keep in touch and be aware of overlaps
... and work together to manage
... I talked aobut that with Tantek explicitly
... 4th topic is federation
... they didn't get to it yet

shepazu: I have thought about this issue
... and that tantek had these reservations
... just 'cause something could be construed as annotation doesn't mean it has to be
... i could do a drop-down in html in 15 ways
... i can characterize data in so many different ways

<fjh> discussed with Tantek, we might consider interop among the different approaches, something to think about

shepazu: that we shouldn't be concerned about duplication
... say there's a complicated way to do +1 with provenance, a whole data model

<fjh> noted that social web has a bunch of material that could be considered provenance, another thing to look at

shepazu: as long as there's a mapping between the simple social-web +1 and the web annotations +1
... then we're OK

<fjh> +1

shepazu: or decompose web anno into social web +1
... that gives us flexiblity of having complex things and simple things and serializeing to a single data model

fjh: that makes perfect sense
... as long as we can interop we're OK
... we should look at the provenence stuff to make sure that mapping makes sense, see what can be shared, etc

azaroth: so long as we can do the transformations, that's great

<fjh> we agree that we should be able to define the mapping between web annotation model and social web

fjh: that's it for this topic
... they are meeting at TPAC, not clear if there's a july meeting

shepazu: can I say one more thing?
... the only problem is that if we have two competing complex data models
... then the mapping becomes more complex

<fjh> I believe social web intent is simplicity, adoption

shepazu: if one of them is simple, then the mapping is pretty clear
... you can always go from simple to more complex

<fjh> we should continue to talk with social web wg

fjh: yes, they're trying to be simple and that's good

<fjh> benjamin and randall may have more to say about social web F2F later, they are at the meeting

Status of Protocol and RangeFinder Work

azaroth: want to go over status of work on protocol and rangefinder
... hurry up, Sanderson, make the changes ;)
... I will try to get the protocol changes added in the next couple of weeks
... I won't make the next couple of calls, should be ready for April 8
... Rangefinder
... last week we talked about having a summary, that's also outstanding from Doug

shepazu: I have substantial changes in pipe for rangefinder, won't be able to do until the weekend
... in the meantime, if people have more to say
... I don't that rangefinder is the only spec we'll do around robust anchoring
... so if folks have other ideas, please put them out
... I'm open to other approaches
... also feel free to put requirements to wiki or ML

Data Model Issues

azaroth: would be good to make some progress on data model issues
... Luc has thought about provenance ontology
... let's skip that until he can come up with a proposal
... from issue list on github
... first issue is about the classes of resource in model should have namespaces removed in JSON-LD context
... the q is about serialization
... people are confused by namespaces and multiple ontologies
... it's possible using json-ld to say that this string means full URL for ontology
... could simplify to just annotation

<paoloC> do you mean this? "@type": "oa:Annotation",

azaroth: could make it easier to understand
... comments?

paoloC: do you mean that the type that is now oa:Annotation would now be Annotation
... when you have to transform to RDF, you have to use base?

azaroth: [something hard to minute]

<azaroth> https://github.com/w3c/web-annotation/issues/12

paoloC: then I'm all for it

<azaroth> {"@type" : "Annotation"} instead of {"@type": "oa:Annotation"}

<azaroth> And have "Annotation" defined as "oa:Annotation" in the context

azaroth: I'm in favor as well

paoloC: there's one example in JSON-LD spec that might be appropriate

<paoloC> "@context": { ... "Restaurant": "http://schema.org/Restaurant", "Brewery": "http://schema.org/Brewery" }

paoloC: is that what you mean? type definition in schema?

<paoloC> +1

<tilgovi> I'm in favor. Just commented on the GH issue.

azaroth: three plus ones and a bunch of plus zeros ;)

<tilgovi> I think a strong supporting argument is that we have motivations rather than subclasses which restricts the domain of classes we need to list in the context to a very manageable set.

azaroth: next easy one is image vs image

<fjh> azaroth will update issue 12 with resolution

<shepazu> https://www.w3.org/2002/09/wbs/73180/annotation-tpac_2015/

azaroth: vocab defines image as a class with two children, still and moving
... so should we use still image and moving image instead of image and moving image

<azaroth> https://github.com/w3c/web-annotation/issues/18

<azaroth> :)

azaroth: I think it's reasonable
... it would be clearer
... any other comments?

paoloC: I was reading the defs

<azaroth> Proposal: Replace dctypes:Image recommendation with dctypes:StillImage

<tilgovi> does it matter much?

paoloC: if we use image, that is also including videos?

azaroth: that's the objection

<tilgovi> I suppose it does if some use case needs to get only one type of thing

paoloC: that was a problem for me
... is there something better than this?

<azaroth> tilgovi: Yep.

<tilgovi> But might many use cases not care and be able to make an intelligent decision about how to display it based on the content type when fetched?

paoloC: is schema ???
... they have mediaobject with various types

<tilgovi> http://schema.org/MediaObject

paoloC: data download and music video object
... audio image and video seem pretty clear

azaroth: do they have a text as well?

<paoloC> http://schema.org/ImageObject

<tilgovi> http://schema.org/CreativeWork is the most specific you can get without getting into particular text formats

paoloC: maybe it's not a good idea
... if you go up to createdwork

<tilgovi> such as Article or Comment

paoloC: some of them are text, but some get really complicated

<tilgovi> It's a little strange. As though text isn't a medium.

<shepazu> http://schema.org/VideoObject

azaroth: any other comments?

shepazu: are we committed to dublin core?

azaroth: I don't think so
... it was convenient
... and they have a reasonably high-level set of classes

shepazu: I have my problems with schema.org, but I have problems with dublin core

<paoloC> +1 for discussing on mailing list

azaroth: two options, one to replace image with stillimage, other to replace all types with schema.org types

<tilgovi> do we have a hard requirement to use either or is this just for putting richer examples with extra properties not directly related to annotation?

<azaroth> http://www.w3.org/TR/annotation-model/#body-and-target-classes

fjh: will there be discussion on ML about Dublin Core? Want to be clear what we're discussing

azaroth: a discussion of what the recommended classes should be
... initial proposal was to replace image with stillimage for clarity

<tilgovi> ah

azaroth: other proposal was to use schema.org classes instead of dublin core classes

<tilgovi> "SHOULD have a class which describes the general content type"

shepazu: and schema.org is active, DC not so active

azaroth: we have five minutes
... 2 big topics left
... we can start thinking about them
... concern when anno has multiple bodies and multiple motivations, cannot align motivations with bodies
... one comment body and two tags, and two motivations of commenting and tagging, you can't tell
... to do that requires a more complex model
... where you have to have some specific resource between annotation and body
... given we want to allow for ??? annotations, that would be problematic

<azaroth> https://github.com/w3c/web-annotation/issues/11

<azaroth> http://www.w3.org/TR/annotation-model/#specifiers-and-specific-resources

azaroth: in order to have per-target or per-body role, would need to be attached to intermediary node

paoloC: I faced this months ago
... I have comments and tags in the same tag
... should I motivate it as commenting and tagging?
... my pragmatic conclusion is that when I read, I don't rely on motivation
... if you have multiple bodies, thinking of adding many motivations is a misuse
... the motivation should explain the main idea of the annotation
... main motivation could be commenting, but you can still add tags

Adjourn

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2009-03-02 03:52:20 $