Data Activity Coordination Group Monthly

Round the table

ivan: We had a pretty good TPAC meeting for CSVW. Made a plan for hwo to move forward
... which meant cutting back on what we can achieve
... mappings to JSON and RDF will be relatively simple.
... we played for a long time with doing something more complicated, a template language etc.
... testing etc would go way beyond that the group can achieve
... so bunch of disucssion going on now
... Interesting point - the charter says we should make a simple mapping to XML, but there's no one in the Wg to do it.
... can't find anyone outside the WG whi wants to do it either. Norm Walsh *might* be interested though
... no interest -> no standard to XML
... and the world won't colleapse

Kerstin: On HCLS
... Trying to participate in the group but don't know bigger picture
... last 2 meetings joined with HL7 with David Booth. he's showing how to use RDF as an enabler
... HL7 getting their act together on RDF - very good news
... brinbging people in from Pharma
... meeteing yesterday had participates from ?? connected with FDA and pharma users/industry
... have been working on core standards for clinical trials data
... really interesting from diff perspectives, not just HL7
... Fire breaking away from XML, but all of these diff strands. Eric P remains key person

phila: Does that chime with work in DC, David?

davidwood: We should pay attention to Kerstin. We have probs and tech solutions but not necessarily political cover
... technically there's a tremendous potential

tbaker: On Data Shapes
... There are groups that are jockying for position in the group
... our concern is that the Cultural Heritage community is represented
... a lot of the UCs are in other areas

<Kerstin> Exxellent ongoing webinar serie on the HL7, W3C work http://semanticweb.com/webinar-yosemite-project-part-1-rdf-roadmap-healthcare-information-interoperability-video_b44757 with David Booth

tbaker: We're happy that the WG is paying attention to the DB that was developed in the DCMI context
... extended to cover areas other than GLAM
... so hope that will continue to be useful

Arnaud: On Data Shapes - Tom's accurate
... I've heard people complain that W3C is too far ahead. here we have the opposite. Wait too long and people are too entrenched
... we're focussing on UCR
... TPAC went well. We had a conversation...
... overheard "tomorrow is when strangulation begins" - but actually it went well
... diff backgrounds and interests in the tech

<Kerstin> Attendes at the HCLS/HL7 meeting yesterday from UCB (Tim Williams) also enangade also in FDA/PhUSE Semantic technlogy project http://www.phusewiki.org/wiki/index.php?title=Semantic_Technology

Arnaud: people looking for different things. Validation cf. Description
... description to do things liike generate the UI - the needs are quite different

tbaker: I heard that the meeting went well too.

Arnaud: On LDP
... main spec is now at PR - just waiting for meeting with W3M
... meanwhile np. implementations has increased. Lots of support

<Kerstin> Also a new attendee, Ingeborg Holt, working for Pinnacle/Open CDISC http://www.pinnacle21.net/ working very close with FDA CDER

Arnaud: we split the spec so that LDP Paging is separate. That's now at CR after 2 LCs - and interest has decreased. Might be stuck in CR for a while
... most of the discussion is about LDP II
... question what are going to do
... we're getting to end of the 6 month extended charter

<Kerstin> Fast Healthcare Interoperability Resources (FHIR, pronounced "Fire") http://wiki.hl7.org/index.php?title=FHIR

Arnaud: making progress on LD PATCH Format
... Everybody wanted it but people couldn't agree on what was needed. I wanted to focus on the main spec
... recent F2F people seemed to have an epihany and agree on an approach
... 3 people took the task of drafting a solution
... but I;m worried it will be a repeat of paging wtih declining interest - hoping to get to LC/CR next week

ivan: The annotation WG has begun to discuss whether LDP might be a good solution for the protocol work
... the annotation WG uses RDF as a model
... the only point is that we try and hide that a little by only talking about JSON and JSON-LD
... so can LDP use JSON-LD everywhere or are there places where Turtle is required

Arnaud: You can use JSON-LD everywhere
... when we began, using Turtle everywhere looked forweard thinking, it's more recently that JSON-LD came along but we've gone back to add it

ivan: TNhe Annotation WG is cheering
... On The Annotation WG
... It started a few weeks ago
... starts withb Open Annotation Model from a CG
... that model is RDRF-based with Turtle everywhere
... but annotation works itself has to and wants to go beyond that so we get into how to store annotations separately, a JS API
... and some ways of defining and querying of annotations (URIs where an annotation is attached)
... so WG is a mixture of people with differnet backgrouns
... some publication people (hence DPub involvement); some Web App devs
... adding annos to data and some SemWeb people
... getting these people to work together isn't easy
... lots of antibodies to RDF in general. JSON-LD is v helpful
... FPWD of the annotation model should be out in the nexct few weeks. Diff between that and CG is that everything is written as JSON-LD wth a context file at the end of the doc

<Arnaud> +q

Arnaud: I;m also co-chair of the Social Web WG
... at TPAC we had the meeting between the two WGs

<davidwood> JSON-LD is great, but please ensure that RDF is still the data model. You might consider using similar language to the way we dealt with that in the JSON-LD spec.

Arnaud: I was surprised by how much similarity there is between the 2 WGs (or wants to do)
... It became clear that the tech is not that different. Posting an annotation is not so differnet to posting a comment on a blog post
... the fact that liaison between the two hasn't been more striongly identified is unfortunate
... could have been brought up earlier
... joint meeiting achieved the goal

ivan: True. Both groups are new, both same age
... I don't know how closely the social WG talks to the LDP WG except via the common co-chair

Arnaud: In the social WG there is an overlap in membership: me, Sandro, Erik W etc.
... makae it easy to see possible alignment
... but it was more surprising that there was zero overlap with annotation WG so we have to identify a liaison person

ivan: Different backgrounds..

Sebastian: On his overview
... We received a lot of new Horizon 2020 projects and I;m driving the DBPedia Association
... right now W3C is doing standards but we also need infratructure
... so for the Web of data, it's unreliable. They don't know how to find the data that they need
... the task, with W3C Help and conferences etc. is to have a reliable Web of data
... standards is one thing. but you need an organisation support
... to make sure standards are applied and relaible
... more things that can't be achieved by standards, like policies
... we need differnet domains, eg law
... you can have public data for law text
... thes emight be annotated by legal experts
... but quality depends on reputation of the lawyer
... quality standards can onkly be defined by domain experts
... so there's some non-tech standardisation necessary. Not sure this is W3C's role

phila: If not W3C, who? DBPedia Association?

Sebastian: DBPA is in a good position to do this. We have the research institute, industry... we might drive this, not sure

Kerstin: I think this is a key topic
... IMI projects in Europe etc. looking at this
... this is eactly the discussion we're having

<Sebastian> +Sebastian

Kerstin: it's a huge topic - commerical money, EU funding, other?

<Sebastian> +q

<Kerstin> This is a key topic also in the pharma industry e.g. for things such as IMI Open PHACTS

eINFRA

Sebastian: SEMANTiCS will be in Vienna in Sept 2015 - maybe have a track on this topic

<Kerstin> Exeample on an joint work between IMI Open PHACTS and W3C HLCS for Datasets and Linksets descriptions http://www.slideshare.net/alasdair_gray/dataset-descriptions-in-open-phacts-and-hcls

Spatial

http://www.w3.org/2014/spatial/charter

LDP II

Arnaud: Not sure that we have a plan quite yet
... but as we went through the dev of LDP I there were features people wanted but we dodn't have time for. So we developed a wishlist. My tip for a WG is to develop a wishlist early on
... that helps people see that their topic is recorded and we might get to it
... If you get to the point we're at now we can go back to the wishlist and work through it, sorting the prioroties
... talking about drafting the charter based on that
... One thing still being discussed - should we charter the WG right away or after 6 month pause
... to allow experiementation
... but the charter schedule should allow WG to work on UCR for first 6 months
... talking about optimisation
... we have the concept of a container
... to explore the info we have to get the container and then GET each element
... want a way to optimise this to get the container and all its resources
... we call this 'Inlining'
... also interest in developing some kind of filtering/querying - but not reinventing SPARQL
... want to be able to select the kind of info that you want. If a container has thousands of resources, how do you select some
... also thinking about communicating changes to the resources
... some way to communicate to the client that something has changed - what? Without getting everything.
... Can we come up woith a way to do this
... We also have Access Control. We put it aside in the first charter. We developed a UCR for that already. That helped a lot
... we could define sometehing that's useful without solving the bigger Access Control problem
... Regarding optimisation - people want to be able to create multiple resources at once
... One thing we talked about - rather than having another big spec is to have smaller specs defining diff mechanisms => LDP becomes a framework of diff technologies
... downside is that end up with less then ideal interop
... but def an interest in this
... If you look at LPD 1 we have conatiners and resources - you can imagine it could have been containers were defined independently. We may want new kinds of containers

Future Future

http://www.w3.org/2014/Talks/1105_phila_semwebpro/#(11)

<Sebastian> SEMANTiCS is international, just happens to be in German speaking countries

phila: Gives quick overview of stategic plans for 2015 re supply chain data and research data