See also: IRC log
<trackbot> Date: 30 September 2015
TPAC registration and hotel reminder: https://lists.w3.org/Archives/Public/public-annotation/2015Sep/0197.html
publishing moratoria: https://lists.w3.org/Archives/Public/public-annotation/2015Sep/0210.html
fjh: starts week of 21
shepazu: we have some time to publish
RESOLUTION: Minutes from 16 September approved, http://www.w3.org/2015/09/16-annotation-minutes.html
fjh: we need to arrange joint meetings
Social Web WG Joint Meeting, TAG Joint Meeting, WebApps Joint Meeting
shepazu: primary purpose of TPAC is joint
meeings with other WG
... need to prioritize around this
... Social Web has decided not to meet but there will be people there
ivan: Digital Publishing Group would like to
have joint meeting, to update on projects and interest in annotation
... Tzviya Siegman is supposed to send email today to arrange meeting
fjh: what should we do with Social Web,
common vocabulary reuse
... schema.org and dublin core conversation?
ivan: they won't be at TPAC
fjh: can start thread re social web people
at TPAC to arrange topics
... think TAG is reasonable, will need to give material head, weren't we
going to summarize the issue
... talk about Rangefinder with WebApps, ok doug
shepazu: yes
fjh: any other joint meetings?
shepazu: social web, webapps, internationalization, digpub, webcsv
ivan: csv not meeting, practically done
shepazu: break out on plenary
fjh: yes, we should have one
shepazu: to let others know where we are, obtain new interest
fjh: yes
... who can do this?
TimCole: where will we be by the time of
TPAC, revision of Data Model in progress
... subgroup to work on validation of approaches to see if annotations
map well against data model
... suspended since model in flux
... would be good to have a stable data model at TPAC
PaoloCiccarese: have had email discussions,
still a few things to sort out
... model has been updated and looks good but more work to do including
reviewing vocabularies
... have told Rob can help
... need to check with Rob on next steps
ivan: even if not perfect, should be able to
discuss editors draft?
... at TPAC, last week of October
fjh: expect we will have published new draft in advance of TPAC, may have editors draft with updates, can plan that will have some more changes for WD later, but can discuss at TPAC
shepazu: have more we can talk about since
last year, Robust Anchoring, new use cases etc so more done than appears
... can talk about copy-edit use case for example, background for data
model changes, worth sharing
... can also talk about what is happening in industry, e.g. generalized
Genius etc
... plenty to talk about
+1
TimCole: agree with that. to prepare for TPAC need to write up use cases and informal documents to explain and illustrate role
fjh: topics for f2F that aren't joint -
Client API and what to do what, Implementation review, HTML issues,
Model and List
... draft agenda at wiki, please edit
https://www.w3.org/annotation/wiki/Meetings#Meeting_Agendas_and_Minutes
shepazu: have talked with people about RangeFinder to validate
<shepazu> http://w3c.github.io/findtext/
shepazu: nature of design
... originally to find any DOM range, including images etc, then
simplified to text
... now FindText API
... more general RangeFinder might happen in future
... small changes to spec in addition to name change
... left old RangeFinder spec in place for historical reference
... (1) internationalization and (2) return value are important
... return value is now a Promise (before object with index to where in
doc and weighting value)
... now that score is up to the application
... discussed with Hypothes.is people and decided no objective score
that can be universal
... Promises alternative to callbacks, went from syncronous API to async
in keeping with other APIs
... enables chaining of processing
... enables iterations to find best match
<ivan> is a general intro to promises
<ivan> Gödel
shepazu: feedback from i4n people re case
folding and normalization (combining character
... as in Gödel, can be written in various ways such as Goedel etc
... Doug lists many ways
... need to say that it is the same string, via normalization
... was saying there is one way to do it
... however different normalization schemes for different purposes say
i4n people
... have updated their spec
<Jacob> Honestly, this is the kind of string reconciliation problem that semantic web should excel at.
<ivan> character normalization draft
http://www.w3.org/TR/charmod-norm/
scribe: what are next steps?
fjh: are all changes in
shepazu: would like to make two changes
... write out algorithm
... have it in prose, need to be detailed
... for searching and edit-distance
... need to update examples to use Promises
<scribe> ... new to Promises but found help, someone from JQuery foundation, should get this in next two weeks.
fjh: I am hearing that we won't have FPWD before TPAC, so maybe we should get feedback at TPAC then publish after
shepazu: big priority for me, will try to get done in next week
fjh: could do FPWD before TPAC, get better visibility, if not we can get feedback at TPAC
shepazu: will work to have update for next week so we can see if we can publish before TPAC
fjh: this sounds good
shepazu: will speak on web annotations at
lecture series at U of Baltimore for library community, working on
presentation
... would like feedback
... FindText will submit talk to O'Reilly March Fluent conference,
FindText API is of interest
... submitted yesterday
http://conferences.oreilly.com/fluent-javascript-html-ca
http://w3c.github.io/web-annotation/model/wd/RequireSpecificResource.html
<TimCole> less consistent
fjh: I think we have consensus on the list for the simpler appearing alternative , proposal 2, despite being less consistent in the model
TimCole: we seem to have consensus but need
to work this issue of requiring specific resource
... adds layer of indirection
... emerging consensus is to allow role on embedded text as well
... Ivan summarized on his post
... asked how to handle multiplicity
<Jacob> really more of an operator than a role
TimCole: example is different languages
<Jacob> role on composite sounds ok but role on choice sounds weird
TimCole: role on group as well as individual
constituents, we need both, does that break the pattern further
... also question of whether we need a subclass
<Jacob> ditto for list
TimCole: for different types of selectors, probably do
<ivan> Ivan's comment
TimCole: suggest we discuss next week after list discussion
fjh: thanks for creating this issue document, it is a great help, we should discuss and hopefully resolve next week
shepazu: regarding choices etc need to better understand the use case, are these multiplicity use cases written up?
TimCole: discussed in community group
... example is choice for different languages for the user to choose
... comes up in selectors, need to get range on video, order matter
... segment of video in region of video, two selectors, timestamp and
pixel region
... will look up and share use cases from Community Group
shepazu: would be useful to make use cases expliicit , not all lists may need same mechanims, eg. choice of language versus selectors
fjh: none
thanks for excellent call