See also: IRC log
<TimCole> scribenick: TimCole
azaroth: Agenda will cover newly opened
issues, including privacy review, selector media table, etc.
... other topics?
... hearing none, any announcements?
... hearing none, move on to
minutes review.
<azaroth> PROPOSED RESOLUTION: Minutes of the previous call are approved: https://www.w3.org/2016/04/08-annotation-minutes.html
azaroth: any objections to minutes?
RESOLUTION: Minutes of the previous call are approved: https://www.w3.org/2016/04/08-annotation-minutes.html
<azaroth> PING review: https://github.com/w3c/web-annotation/issues/204
azaroth: Issue #204, results of
Privacy review.
... summary - two most important comments, 1) request to
recommend https
... 2) there should be some way to opt out of being
annotated
... similar to how robots.txt requests crawlers not to
crawl
<dwhly> +q
azaroth: reality is there is no
way to preclude annotating, but you can request
... Privacy also asked about being able to delete, etc., but
this is supported.
dwhly: okay with https, but lots
to say about 2nd suggestion
... we have been doing research, outreach, interviews, with
people across the spectrum
... of project, users, blog posters, etc.
... there is a diversity of opinion, but some consensus that
there should be some way page owners can signal would-be
annotators
... would be interested in the opinion of the WG whether they
thought this was within their remit
... there are a wide variety of things that might be
signaled
<shepazu> (privately/locally annotated vs published)
dwhly: proxied or not, annotated
or not, annotated only if annotations are moderated, etc. -
long list
... mechanism not clear yet, but first question is should a
signaling mechanism be considered.
<bigbluehat> "annotation service" == "publisher" ?
dwhly: would require some negotiation, but annotators would need to be able to override if in public interest
<tbdinesh> first. how to we identify a page owner? (so as to signal)
dwhly: these are some thoughts within hypothes.is, interested in knowing what others in the WG think.
azaroth: Could you (Dan) write-up a summary of data gathering so far and post to WG list serv?
dwhly: yes, could be a couple of weeks
shepazu: a big problem on the
Web, there should be a place for standards here
... there are useful things beyond if you want to be annotated
or not
<ShaneM> Regrets for today - sorry!
shepazu: e.g., do I have my own
annotation server?
... so there are degrees beyond do I wanted to be annotated at
all?
... what are the other useful end points (annotation
repositories)
... maybe only be able to annotate to curated sites?
paoloCiccarese: sounds cool, but what's the goal?
paoloCiccarese: are we just trying to publish guidelines?
<fjh_> +1 to paolo
shepazu: yes, guidelines
<bigbluehat> can we untangle the terms and vectors we're discussing?
shepazu: a way to express intent
paoloCiccarese: examples - Canadian site that allows people to make comments. Don't like but how can you block?
<bigbluehat> do-not-blog-about-me.txt
paoloCiccarese: history is that blogs ultimately will do what they want.
shepazu: distinction between can you annotate and can you publish annotations to a site and should annotations be displayed in the context of my page
azaroth: to Doug's point about
where to annotate - there is a way to say in the model [scribe error, corrected later - should say 'in the protocol spec'] that
this is my preferred annotation service
... also agree is provide a way for content publishers to
express preferences, can't enforce in this kind of distributed
system
... we might want to have an initial list of what content
providers would like to be able to say
... then we can work through these and decide what's reasonable
for first draft
dwhly: robots.txt is not entirely
a good analog for what we do.
... robots.txt is respected by some but not others
... that will happen here as well
<azaroth> +1 to Dan. If the big systems respect it, then the damage is limited. Robots.txt limits bandwidth damage, this would limit social damage
dwhly: but would be good to
provide a standard way to express preferences and formalizes an
ontology of these preferences
... this is social signaling basically
... which is a benefit
paoloCiccarese: what are the
connections to legal issues
... does any of this fall into legal space?
<Zakim> shepazu, you wanted to clarify signally in protocol
shepazu:W3C doesn't do a lot around legal, but there are intersections
shepazu: there are legal
precedents, lots that happens in legal space around W3C
standards
... I would hope that the markets would coalesce around our
ontology and avoid the need to turn everthing into legal
paoloCiccarese: so this assumes that we know that what we propose would not contradict legal.
<bigbluehat> does proxy + overlaid annotation == publication (for instance)
shepazu: what we have now happens
in HTTP headers, but maybe we also need to think about putting
something in the HTML
... and the right reference for what we have currently is in
protocol not model
<dwhly> +1 for collaboration!
tillgovi: are these issues also
being dealt with elsewhere
... some of what we're talking about is not just about
annotation
... e.g., annotation doesn't require displaying as overlay, and
content can be talked about beyond the scope of annotation
dwhly: in discussing with a
variety of folks about signaling
... where it might be
... an immediate concern is that I may not have access to all
of these mechanisms (headers, etc.)
... could it be signaled at the bottom of the blog?
<Zakim> tbdinesh, you wanted to mention pages dont have an owner who can be identified by a community
dwhly: concern might be that someone could spam the intention of the page owner.
tbdinesh: page owner could lots of different kinds of agents
<Zakim> azaroth, you wanted to suggest Social Web WG
tbdinesh: our model doesn't really say how the page owner is identified
<bigbluehat> +1 to azaroth re: social web
azaroth: this falls into the
remit also of the Social Web WG
... would be good to work with them
<shepazu> +1 to working on this outside Web Annotation WG
<bigbluehat> note their charter expires this year also...
azaroth: re Dan's and Dinesh's questions - if we have a way for blog authors to signal intent, I would hope that platforms would allow for that
<shepazu> bigbluehat, yes, might be something outside either WG
azaroth: the author then is the person that controls that particular preference
<bigbluehat> new WG most likely
tbdinesh: we are getting into
question of ownership of pages on the Web
... we are asking the page owner to identify themselves some
how
<azaroth> scribenick: azaroth
TimCole: In terms of next steps,
don't want to get derailed, but do want to respond to the
feedback. What modifications do we make to the specs?
... What would satisfy the issue raised, and what could we
leave for additional work?
<scribe> scribenick: TimCole
<dwhly> I don't think the page owner needs to identify their real-name, they just need to state a preference in some agreed upon location.
azaroth: any objections to recommending Https?
<bigbluehat> not to recommending. objection to *requiring*
<azaroth> PROPOSED RESOLUTION: Recommend HTTPS in the protocol doc
+1
<azaroth> +1
<tbdinesh> +1
<shepazu> +1
<takeshi> +1
<dwhly> +1
<PaoloCiccarese> +1
<bjdmeest> +1
<bigbluehat> +1
RESOLUTION: Recommend HTTPS in the protocol doc
azaroth: for signaling, if we can
come up with list, hopefully based on interviewing Dan has been
doing
... and then once we have that list come up with a
mechanism
... goal would be have list by F2F
dwhly: yes, we can summarize
before then what we've heard
... and then have a discussion at f2f
<tbdinesh> dwhly: to state a preference on *a* page requires a lot of thought. easier as post-event-signalling. maybe an abuse signalling?
azaroth: will take action to talk
to social Web WG to see if they have thoughts
... to see if can come up with around social publishing rather
than just annotating
... it's about making it public.
shepazu: agree with sentiment that this is a larger problem, here's what we can do now.
<tbdinesh> dwhly: also a blog is generated (owned) by an institute generally but an owner owns the content in a way of a/their blog post
azaroth: any further next steps
on this one?
... hearing none, we move on
... we should add a privacy consideration section
... rest of the privacy feedback looks like editorial. Any
contrary thoughts?
<azaroth> Media Table: http://w3c.github.io/web-annotation/model/medias.html
TimCole: this is mostly informational people should review this and see if they agree and if adequate for helping with testing
azaroth: actions here are for the WG to review and post comments to the issue #203
<azaroth> github link: https://github.com/w3c/web-annotation/issues/203
azaroth: for next week let's try
to call it done unless objections are raised between now and
then
... any other thoughts about #203?
... to clarify, with regard to conformance we don't want to say
that an implementation must support all selector types
... this table is a way of saying what we think an
implementation supporting a media type should support.
<azaroth> github link: https://github.com/w3c/web-annotation/issues/200
<shepazu> 200 = Should oa:start / oa:end be xsd:nonNegativeInteger or xsd:integer ?
azaroth: should be may explicit
where integer values should not be negative (right now just
says any integer)?
... a question of specificity
<shepazu> Deref of namespace should go to vocab? #199
azaroth: any thoughts or
objections on #200.
... stian suggested putting
type into the json-ld context, but this can sometimes make a
mess in the json instance
... stian and azaroth will test and offer closure, subject to
review of the WG
+1 for specifying non-negative
<azaroth> PROPOSED RESOLUTION: Accept xsd:nonNegativeInteger for start and end in ontology and vocab
<azaroth> +1
+1
<tilgovi> +1
<bjdmeest> +1
<shepazu> +1
RESOLUTION: Accept xsd:nonNegativeInteger for start and end in ontology and vocab
<azaroth> Github Link: https://github.com/w3c/web-annotation/issues/199
azaroth: #199 is about what
should happen when you dereference the namespace
... when you dereference a namespace you typically get
documentation about the vocabulary
... there are multiple options, a landing page with link to
human documentation and machine-readable, you could get turtle,
etc.
... we will discuss next week and see if we can
close/resolve
... azaroth will try to make concrete proposal
shepazu: Shane will help put
together the framework for general json-ld testing and will
work with shepazu
... hopefully will be able talk about next week.