See also: IRC log
<trackbot> Date: 27 April 2015
<tzviya> agenda https://lists.w3.org/Archives/Public/public-digipub-ig/2015Apr/0098.html
<tzviya> http://www.w3.org/2015/04/20-dpub-minutes.html
tzviya: let's look at last week's minutes
... motion to accept?
... minutes approved.
... today's agenda
<tzviya> https://lists.w3.org/Archives/Public/public-digipub-ig/2015Apr/0098.html
pkra: first look at survey
... 34 responses
... sent to 93 people
... ok result, but not too exciting
... people don't feel qualified to answer
... good coverage on all questions
... 37 questions total
... there's an unsurprising bias towards math
... partly due to me leading the survey
... lots of people talked about mathml
... it wasn't a random sample
... run through a few of the questions
... first section was about background
... bias towards CS and math
... 2nd was professional background
... didn't have a lot of aggregators
... most people were researchers
... researchers were primary audience
... we forgot to ask about students as audience
... which platforms people serve
... a few comments pointed to "the web is our platform"
... most people were focusing on desktop
... on the low end was print and ebooks
... next section was about content
... subject domains
... the question about reusing content was unclear
... "were people actively reusing content"
... prev. q was about making content resuable
... so people just repeated answer
... there was q about standardization
... there will be fun quotes from that
<pkra> https://www.w3.org/2002/09/wbs/64149/DPUB-STEM-2014-12/results
pkra: next section about authoring
... there were a couple of problems
... first question didn't get responses we wanted
... should have asked explicitly about STEM fragments
... want to know why people lose information when converting
... but questions were too vague
... people talked about transformation of text formats
... surprised to hear people do version control, did mention things like
git
... "we see more and more JSON in scholarly publishing"
... that was unexpected answer to STEM fragment storage question
... and very positive
ivan: are some formats other than MathML that are widely used?
pkra: one question asked that
... not at this point
... alas, I don't have all the data in my head :)
tzviya: question 21, maybe?
pkra: yes, 21
... CML
... there isn't that much
... someone from Wolfram had CDS
... not as much as I'd hoped
... in the authoring section
... Q15, how do people provide access to content fragments
... that was a good result
... people are used to using XML
... next section on delivery
... how is content delivered
... HTML was ahead of PDF
... 32% to 26%
... most produce both
... there was an open question about desired methods
... with ten longer comments
... Q about migrating away from PDF
... people said Yes
... not enthusiastic about PDF but said they have to
... Q about exposing data on web
... answers were all over the place
... Q about embedding scientific data via attributes, microdata, etc.
... quite a bit of stuff there
tzviya: more summary would be good
pkra: we didn't mention stem fragments again
when asking about bridging reading and authoring
... lots of fun quotes about reading
... the workflow section was tricky to write
... similarly solid in results, no surprises but good data
... PDF comes up because we have to
... a11y section I haven't looked at much
... that was a challenging section
... more answers about people not having expertise or answers
tzviya: I have a few questions
... first
... everyone should read through this even if STEM isn't your bag
... thanks PKRA!
... so what do we do next? There's a lot of information here, and it's
all over the place
... might be helpful to focus on a few points and go from there
pkra: one challenge is to extract data in
efficient way to spreadsheet
... yes, it's a lot, and it's not clear what to focus on
... a meeting of task force will help
... to get an in-depth summary
... original idea was to publish a note by the IG
... but the input in some sections is not viable
... for example, a11y is not balanced enough to provide good feedback
tzviya: if we try to write a note summarizing everything, it would be huge
pkra: would be good to make data available in anonymized fashion
ivan: I've already produced a spreadsheet
... so I can just remove a column
pkra: yes, but maybe that's a different
conversation
... I had trouble with the spreadsheet
... Ivan, is it legally possible to publish data?
ivan: if it's anonymous, then it's not a problem
<astearns> not just names, but any identifying details in free-form responses
pkra: then the note can be much more focused
... it's all anecdotal
Bill_Kasdorf: while the anon. data are
useful, what's most important are the messages
... sometimes they're clear and sometimes they're contradictory
... don't need to cover all issues equally
... what did we hear that was notable?
pkra: I agree
tzviya: OK. Any other comments?
ivan: You also need people around you?
pkra: I will rely on the task force, but more people will be great.
ivan: if anyone in the group has some experience in managing survey results, volunteer!
pkra: I'm also not experienced
NickRuffilo: I've done this
... if you have any questions about creating non-leading questions or
analyzing results, please let me know
... I'll pull up articles I've written about this
tzviya: NickRuffilo is our new favorite
Bill_Kasdorf: too much reliance on statistical analysis would be suspect due to the small and biased sample
Bill_Kasdorf: more an editorial task
<NickRuffilo> Excel TIPS: http://publishingperspectives.com/2013/01/tips-for-technologists-7-excel-with-excel/
Karen: do I have confidence this is not public, just members-only? I don't think raw results should be public
ivan: we can check
pkra: it's always asked me to log in
ivan: it's not public
Karen: OK. If there's a quote, it will not
be attributed?
... more a qualitative report
ivan: it should never be quoted or attributed
pkra: most people would be OK with that
tzviya: we've never asked permission to
quote people
... we have our existing STEM task force, and Nick has been drafted.
<pkra> or otherwise shanghai you ;-)
Karen: we have a replacement AC rep for Copyright Clearance center, who may be interested in STEM
tzviya: thanks Karen.
... moving on...
tzviya: a few weeks ago there was a discussion about fragment identifiers
https://lists.w3.org/Archives/Public/public-annotation/2015Apr/0051.html
tzviya: which ended up on the annotations
list
... Ivan will lead us through this
tzviya: lots of discussion on what makes a legal identifier
ivan: is rob around?
... we discussed that three weeks ago that the model for selectors in
open annotations document
... is in fact a very rich and powerful collection of terms
<tzviya> ivan: perhaps starting with position of the selection model in open annotation is a good
<tzviya> ...not bound to one media type
<tzviya> ...the model is an abstract model, not described in terms of URI
<tzviya> ...perhaps it's possible to turn it into fragment ID
<tzviya> ...then there was further discussion of when a frag ID is legal, etc
<tzviya> Rob's position: https://lists.w3.org/Archives/Public/public-annotation/2015Apr/0054.html
ivan: the fundamental issue is the following
... a fragment identifier is bound to a specific media type
... you must register for each and every media type
... so you can't just declare the OA model for the entire world
... if we go down that route and use the selector model
... the correct way is we define them as part of a URI
... and then we register them for some of the media types for which they
are useful
... html, svg, etc.
... we can do that
... then it can be combined with other mechanisms as it's done with web
packaging
... it's not clear to me who should work on this
... I have the impression that there's some sort of agreement that if we
restrict by media types we can do this
... azaroth, is this a fair summary?
azaroth: yes, that's it
Bill_Kasdorf: keeping in mind the mission of
the various groups
... DPUB is expressing needs, not writing standards
... I have a strong interest in what Anno WG comes up with
... as far as DPUB, our job is to surface the issue and work with the
appropriate WG
... the OA model you recommended came out of a CG
... I'm not trying to wash my hands of it
ivan: I understand and agree
... two comments
... we already do this approach with structural semantics where we're
involved with PF
... the other thing is that we need to recharter this group
... maybe we can then go beyond what we have here
... we'll see where it goes after September
<TimCole> http://www.w3.org/2015/04/22-annotation-minutes.html
TimCole: it was discussed at Anno F2F a bit
... if this is important to DPUB
... you need to push Anno WG
... talked about in context of rangefinder
... but it's not high on our priority list
... might be good to bring this up as collaborative
... the anno model allows lots of selectors
... not all will work as fragment identifiers
... also, EPUB already has something that kind of works
ivan: I agree
tzviya: we all are talking about epub cfi as
if it solves the issues, but few people use it
... there are too many options
... it's still a multiple choice question
... i can use xpointer, I can use CFI, but what happens with packaging
and building systems and epubweb
... those urls look like multi-part mime
... so maybe we should get used to those semicolons in URI
ivan: for me, the packaged uris don't look
that funny
... if you combine with packaging, what is done in the packaging spec is
what should be done
<clapierre1> looks like Readium supports the EPUB CFI https://github.com/readium/readium-cfi-js
ivan: we are getting into technical
discussions
... good topic for f2f
... we should have clear and clean idea of pros and cons of CFI
... it's in the same space as the selectors
... CFI provides you with a fragment ID
... is CFI was completely useful then the package spec approach combined
with CFI and we are done
... so we need to have a clear idea in NY whether CFI works, or it does
not work
... if it does not work, we need to look at alternatives
Bill_Kasdorf: don't want to conflict with anno
tzviya: we followed offline with berjon and
michael smith
... HTML will not have a formal proposal for an element, so we should
pursue role with aria
... in the past, some roles have been promoted to element in HTML
... so we're moving forward with ARIA role
... maybe HTML will take it up in the future
TimCole: since most annos are third party
... there might be use cases where people mine footnotes as they mine
annotations
... we should see how footnotes might be transformed into annotations
... we should keep that in mind
tzviya: we agree
... footnotes are somewhere in the middle of content and annotations
<azaroth> +1 :)
tzviya: let us know if you're coming to the f2f
<tzviya> http://www.w3.org/dpub/IG/wiki/May_2015_F2F_Logistics_and_Details
tzviya: see you next week
<ivan> trackbot, end telcon