See also: IRC log
tbaker: We had a half day
workshop on RDF validation, then haf day panel session on same
topic
... one of the big topics of this year's DC conference
(Austin)
... in addition we had a call last night with the DC
Architecture Group, working on reqs for Application
profiles
... EricP was there
<davidwood> DC-2014 conference: http://dcevents.dublincore.org/IntConf/dc-2014
tbaker: Thomas Bosch, Manheim,
has been compiling reqs for validation. Part of them are
related to what we've been talking about
... bu the DB itself is much borader
... It's on the agenda for thje first RDS call (2 hours'
time)
... a few take aways
... the cultural heritage people in the group, many of whom are
aggregators and handle messy data in lots of language
... espeicially Europeana
... is it helpful to frame the problem as one of validation as
a number of them find the notion counter-productive
... if we think in terms of correct/not correct
... what they want is to analyse the data and ...
... the Germand digital library, for eg
... data reports give to the data provider
<davidwood> Tom's DC-2014 Tutorial on RDF Validation in the Cultural Heritage Community: http://dcevents.dublincore.org/IntConf/index/pages/view/rdfAP-Tutorial
tbaker: in which they point out anything that are show stoppers
<davidwood> DC-2014 Special Session on RDF Application Profiles and Tools for Metadata Validation and Quality Control: http://dcevents.dublincore.org/IntConf/index/pages/view/rdfAP
tbaker: but also want to make
suggestions on how the data can be improved
... areas where the data doesn't seem to be of good quality and
how serious that is
... they want to work with the suppliers
... eg if you use URIs in this position it will improve the
linkage etc.
... that was one of the take aways
... some presentations, eg a tutorial looking at SheX, SPIN
etc.
... the reqs database has examples in various syntaxes
<tbaker> http://etherpad.wikimedia.org/p/dcmi-ap-rdf-14-10-2014
tbaker: ^^ is from last night's
call.
... may be a DCMI view, a database view etc.
... we'll be keen from a DC perspective whether the WG takes
this on as a resource
<davidwood> My slides are available at http://www.slideshare.net/3roundstones/w3c-data-shapes-working-group but please note that I shamelessly stole most of the contents from PhilA's blog post at http://www.w3.org/blog/data/2014/09/30/data-shapes-working-group-launched/
davidwood: I put a couple of
links in so you don't need to find them
... most of hte content is from the W3C blog post
<davidwood> The librarians at DC-2014 very clearly see how RDF validation can help their efforts, such as BibFrame adoption, data cleanliness of MARC records, etc. Engaging with that community is really quite important for the Data Shapes WG if you want them to adopt the eventual standard.
davidwood: people like the DPLA,
Kevin Ford from Loc, Richard Wallis OCLC
... Karen Coyle was there
... these people really see RDF validation in relation to
library data very clearly
... they understand where validation will assist them with
Bibframe adoption etc
... the flexibility to make it work etc. MARC records present
specific issues
... the community is receptive to new ways of working if not
tied too tightly
... noting, not fixing, is useful
... if the DAta Shapes WG expects to produce standards that
people want to use, the library community is in the middle of
this - modern cataloguing etc
... but of course standards have to work for them. Need to get
them into the WG somehow
... if the data Activity wants a good win, then there's an
opportunity to do so if we can get the people in
tbaker: The DC community/CH
community really sees this WG as providing the tech standard
but they also want a high level language with a less tech
background to author profiles
... sometehing that can complie down to an underlying
constraint vocab
... but I think the hope is that there will be either as a
result of the W3C activity, or from the DCMI activity, some
sort of higher level language that is easy to use
... in the CH community
... that lets non-experts write application profiles
davidwood: That is the challenge.
The data Shapes WG needs to use those as inputs so that its
outputs are not restricted to easier ones
... the ones that i'm most interested in to drive UI generation
are all very restrictive, but there is a fund of use cases for
warnings, options etc.
... flexibility in general
... and they need to be captured, or there's a risk of it not
providing a tech foundation for DCMI
<ericP> 200+ actually
tbaker: I should add that the
reqs DV has 200+ requirements, many from the Data Shapes
discussion
... broader than just CH community
<ericP> though a lot are from a wholesale inclusion of OWL functional semantics
tbaker: That link to yesterday's
call provides some of the context
... it's Eric and Karen discussing ther reqs data base and its
use
ericP: is here...
davidwood: the world needs more people like Karen Coyle
phila: WDYT? ericP ?
ericP: The prods to fix data are
probably aplication side
... it's likely that people will want something that helps
people fix the problem, not just being told you're wrong
... we'll be doing requirements for 3 months
... so that's going to set the scope
... whether it's MAy SHOULD MUST etc
davidwood: I recall that there
was a list of levels... from DPLA?
... 6 levels, i.e. beyond RFC 2119
... maybe in the Google Doc
tbaker: I think mark's
presentation should be online soon
... it should be in that
... I believe they have a star system
... 1 - 5 star
ericP: The obvious tech approach
is Must/optional/prohibited
... higher than that can be done by compiling down to that
short list, run check once per star
... couple that with intelligent output
... You can still do multiple levels of conformance, even if
you have binary
<davidwood> Rough notes from the DC-2014 panel: https://docs.google.com/document/d/1ExmMZCg5mApOzNAHvL4pDS_geHuoWcj26nNW3vmmiRM/
<davidwood> Note Tom Johnson's comments re levels of conformance.
bhyland: It may be obvious, but are there conduits between library and W3C beyond people like Dave and Tom?
davidwood: we call it Tom
bhyland: But is Tom doing this as
part of the day job?
... We see other areas where people are reinventing wheels
because they don't know about W3C
tbaker: Well there's Eric
... Eric was talking about putting a heirarchy on the reqs
database
ericP: I need to do some social engineering here, yes
ivan: One more name... we got someone from the Loc joining the Annotation WG (Ray Denenberg)
Soeren: Since you were asking
about connections... my team at Fraunhofer are the ones that
work on the german digital library
... establishing use casesa nd connections
... we do al the mappings to Europeana etc.
... I know that validation is a very time consuming job
<ericP> Soeren, any chance you'd have someone to join the validation WG/
<ericP> s/\//?/
<Zakim> davidwood, you wanted to discuss library members
<davidwood> OCLC, Library of Congress are W3C members. It would be good to approach DPLA for membership. OCLC and LC should be invited to participate in the Data Shapes WG.
<davidwood> "Mark A. Matienzo" <mark@dp.la>
<davidwood> ^^ DPLA
<ivan> Emmanuelle Bermès
discussion about inviting DPLA to join the WG
ivan: ^^ Emmanuelle recently went
back to French national library
... the France office had a disucssion with them about poss
mmebership
tbaker: I co-chaired trhe libraryt LD group with her and Isaac
Arnaud: we have an informal call
today to allow people who are considering joining to dial in
today
... we did have a whole list of AC reps who voted in favour of
the WG, saying they would participate but who haven't joined.
So we have been pinging them
deirdrelee?
<deirdrelee> i'm here in google campus, so mic not working
ivan: The CSVW WG has been
reduced to a very small active group
... there have been a number of e-mail threads on converting
CSV files to JSON, RDF, XML etc
... but not yet reached consensus
... all a bit afraid of getting itno something too cmplicated
that we can't finish
... there's a doc on the metadata that is fairly mature but
faiely complex
... we hope we'll be in good shape after TPAC
<deirdrelee> :\ but written report....DWBP going well, 2nd public working draft of UCR published yesterday http://www.w3.org/TR/dwbp-ucr/ Will invite new use-cases from wider community (email to be disseminated in next few days). Will discuss this in TPAC. Also progress on other deliverables
Arnaud: LDP should go to PR next week, probably
Arnaud: I was hoping we'd make
the decision this week but it didn't happen
... but we had a load of late comments... from WG members
... bit disconcerting when WG members finally start paying
attention
... good number of implementations
... one feature, Indirect Containers, that is less well
supported and we're working out what to do
<bhyland> Linked Data Platform 1.0 - Fourth Public Working Draft - third last call (2014-09-16) see http://www.w3.org/TR/2014/WD-ldp-20140916/
Arnaud: LDP Paging spec. No
significant comments. Same kind of comments, doesn't HTTP Range
suffice? Well no, we're working towards CR with that
... And we have a new LDP PATCH doc
... controversial topic. People feel differently about it, hard
work to get agreement on what the requiremetns were
<bhyland> LD Patch Format First Public Working Draft (2014-09-018), see http://www.w3.org/TR/2014/WD-ldpatch-20140918/
Arnaud: doesn't address all
possible cases but does the ones that the WG agreed to
address
... why isn't SPARQL Update enough? Because it has a level of
complexity that not everyone wants
... we have extended the charter to end November - not long
<Soeren> sorry have to drop out (have to catch a plane)
Arnaud: we're in pretty good shape. Now talkinbg about rechartering
Soeren Auer
<ivan> Sören
<bhyland> To German readers, more info on Soren http://www.iais.fraunhofer.de/linkedenterprisedata.html
bhyland: The US has been a little
slow off the dime around publishing open data anda paying
attention to standard but once we get going it's hard to
stop
... what I notice from the many events in the DC Area, ther's
so much happening and people are not well informed about what
other people are doing
... there are Meetups
... Josh Tauberer has a book out about this
... the Data Act passed in May, Prez signed it
... I've given up on individual gov agencies joining W3C but we
need a bridge with people like New York GovLab
... to inform officers on how Web can be used
... they're not going to be in the WG - too much time - but if
we had a profile-raising action
... at workshops, etc. then W3C will have lost their
credibility and opportunity in the US
davidwood: It's clear that
there's a lot of momentum gathering around ipen data for
transparency, following Transparency Act and Data Act
... it's the concern du jour
<bhyland> FYR, Josh Tauberer's Open Government Data book as a resource: https://opengovdata.io/
davidwood: but there's a line
between open data/gov data and W3C
... people talk about this stuff without understanding where
those boundarfies are
<deirdrelee> +1 sounds familiar to discussions within dwbp wg
<bhyland> Josh outlines 14 principles of Open Government Data & people are looking at these … might be helpful to have Josh informed on what W3C can do / Recommendations, WGs / Community Groups, etc.
phila: Talked about nascent plans for workshop in Washington next year around publishing/link with supporting data sets
DavidWood: Recent blog post from
Nature publishing about data and publishing being intimately
related
... average file size is 1.3 MB, it's the complexity that's
hard for them
<bhyland> DavidWood's reference to Nature re: size of datasets & complexity as well suited for RDF approach, see http://blogs.nature.com/scientificdata/2014/10/13/size-doesnt-matter
<bhyland> Title: "Size doesn't matter" 13-Oct-2014 posted by Sarah Payne, Scientific Data Updates blog
DavidWood: that's where W3C has been prepared for a long time.
<bhyland> Next meeting 19-Nov-2014
<bhyland> have fun at TPAC all! We'll miss you!
<deirdrelee> thanks everyone, bye
<bhyland> Thanks PhilA. Interesting call!