DCMI RDF AP Task Group
http://wiki.dublincore.org/index.php/RDF_Application_Profiles
Meeting date: Feb 11, 2016. NOTE NEW TIME: 8am PST, 11am EST, 5pm CET.
Meeting link: https://meetings.webex.com/collabs/#/meetings/detail?uuid=M2O5NEDCS7DKY32GWG8A46DS0D-JV0D&rnd=756673.45093
Meeting minutes: https://etherpad.wikimedia.org/p/dcmi-ap-11-02-2016
Attendees: TomB, Hugo, Antoine, LarsG
Regrets: Valentine, ThomasH
AGENDA:
1. Admin
DCMI technical board would like:
- a "brief statement detailing the next steps for the group and maybe identify a new deliverable to cover the next period of work..
- "formalise the closure of the current work with a brief of 1-2 pages summarising the first outcomes with pointers to the deliverables and wiki". Could be used for dissemination on LODLAM, CODE4Lib...
ACTION: Karen and Antoine to act sending the material that the DCMI tech group requires.
Corey: could it be sent for the group to review?
Karen: sure
2. Alignment with W3C Data Shapes
- W3C activity
Karen: 2nd draft of SHACL is out . Big issue: we are not getting many comments. Need to spread the word. Also need tests, but
test suite format is not clear and draft is still changing, so hard to come up with tests. See proposed simplifcation below. Holger
is an OO programmer. Authored RDF-for-OO-programmers on W3C site. In first iteration, very OO - alot of abstract classes.
Arthur Ryman came on as second author, trying to align more with RDF. With Holger, it was expressed with SHACL not RDFS.
Interesting aspects, but for someone who wanted to develop tests, could not use RDFS tools but SHACL tools (non-existant).
Arthur created RDFS vocabulary for SHACL. Hopefully SHACL is moving closer to standard RDFS; hopefully possible to create
tests with RDFS tools.
Proposed 'simplification' to SHACL: https://www.w3.org/2014/data-shapes/wiki/ISSUE-95:_Metamodel_simplifications#Proposal_1
Hugo: What is meant by RDFS? Because SHACL is defined in RDFS.
Karen: Originally in RDF, but not RDFS.
Hugo: The classes are there though.
Karen: Engine itself could not be defined in RDFS, only SHACL.
Tom: I'm a bit confused
... Is SHACL more than a type of pattern matching spec like SPARQL?
Karen: it is, but it's quite complex
Antoine: this is about meta-modeling, a like
Hugo: it can help specifying the domain and ranges of some propreties - for SHACL designers
Tom: I'm a bit worried by this focus on semantics. It competes with RDFS
Antoine: it's really about metamodeling
[Tom reads some very confusing part of the SHACL spec]
Karen: please send to the group.
Tom: I've discussed with Holger on Twitter
Antoine: this is part is about requiring inferencing or not. Our discussion on what consistutes a graph.
SPARQL does not require the triple store to do the RDFS inference
the RDFS expression of SHACL is something that's more related to the definition of the vocabulary.
A bit like it would have an XML Schema in order to express it in XML
Karen: the same could be happen in JSON-LD
Tom, Karen: [SHACL vs SPQRQL]
Karen: some of SHACL is just wrapping SPARQL functions
... but there are some things people want to do in SHACL that SPARQL doesn't do (well) for example order of properties.
ACTION: Tom Johnson to ask the group about the RDF Ruby RDF core team implementing SHACL
Ruby group discussion noted by W3C group - great excitement!
Tom Johnson's post about doing a Ruby implementation of SHACL got alot of excited attetntion.
...Holger says that TopBraid will make its own (partial) version. If Ruby is the only implementation of the standard....
Antoine: Aren't two implementations required by W3C process?
This is a problem re. W3C requirements (features should be implemented by two implementations)
DCMI repository in W3C Data Shapes Test Suite by Karen:
https://github.com/w3c/data-shapes/tree/gh-pages/data-shapes-test-suite/tests/dcmi
- EDM SHACL activity report (Valentine/Hugo)
Europeana SHACL test file: https://github.com/hugomanguinhas/europeana/tree/master/edm-shapes/src/main/resources/etc/edm/shapes
EDM validation rules: https://docs.google.com/spreadsheets/d/1dI29hgckyHn3B0h5shmbYX3jcp3PZfX7TA30Ohp_dMI/edit
Hiugo: there are things difficult to express, like checking cycles
Karen: I mentioned that there is discussion on DCMI group.
... but nobody in the group is seeing
... it would be good to have some questions passed to the group
... without taking too much time
... do you have some open questions?
Hugo: I have a couple of them
Karen: you can send them to the W3C list.
... even the questions that were discussed w/ Holger
Hugo.: I can send first to the RDF AP
Karen: you could send directly to the group
Antoine asking about the strategy for posting
Corey: this one looks like a regular listserv
https://lists.w3.org/Archives/Public/public-data-shapes-wg/2016Feb/
Karen: it should be a different list
https://lists.w3.org/Archives/Public/public-rdf-shapes/2016Feb/
Karen: this is an old, from the very first workshop
Antoine: W3C practice is different is other groups. see https://www.w3.org/TR/vocab-dqv/
"If you wish to make comments regarding this document, please send them to public-dwbp-comments@w3.org "
Corey: this is unfortunate
Tom: the message in the archive still refers to the workshop
Karen: I can flag this to Arnaud
ACTION: europeana to formalize the open questions that Valentine told us about at https://etherpad.wikimedia.org/p/dcmi-ap-28-01-2016
--ONGOING...
3. New requirements and APs?
Are these two LDP-related requirements in scope for DCMI group? Are they related to Lars' requirement below?
http://w3c.github.io/data-shapes/data-shapes-ucr/#uc31-ldp-post-content-to-container-of-a-certain-shape
http://w3c.github.io/data-shapes/data-shapes-ucr/#uc38-describing-and-validating-ldp
Karen: these have notbeen implemented in SHACL
Corey: these sound like gaps in the LDP
... they're relevant but out-of-scope for SHACL?
Lars: yes
... on the other hand I've talked to Phil
... he suggested that I submit this to Arnaud
... let them figure out how best to implement it
Antoine: I have some (maybe stupid) feedback I'll send it.
Lars: ok
... once feedback is there and the I-D is updated it will be sent to Arnaud, cc DC-ARCHITECTURE
Lars' requirement for http negotiation of shapes
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A1=ind1512&L=DC-ARCHITECTURE#7
Here's the proposal Internet-Draft: https://tools.ietf.org/html/draft-wilde-accept-post-00
LDP definition: https://www.w3.org/TR/ldp/#header-accept-post
- Miika Alonen's AP
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=DC-ARCHITECTURE;c364c4f8.1509
Antoine: this is an old case we never reacted to
... but it was interesting
Karen: it's a nice list of test, could be tests for SHACL
Corey: all the test cases are implemented as Javascript
From their page:
This software performs the validation of DCAT-AP rdf files.Validation is performed via a web form which loads a file (Turtle, RDF/XML, N-triples, JSON-LD) as graph in a triplestore and it queries the triple store with a SPARQL query.The output of the validation can be in XML (HTML with xslt transformation), JSON, Text, CSV, TSV.
https://joinup.ec.europa.eu/software/dcat-ap_validator/description
https://github.com/EmidioStani/dcat-ap_validator/tree/master/tests
https://github.com/EmidioStani/dcat-ap_validator/blob/master/rules/rule-0.rq
It also does error messages
the binding of error messages is a bit a hack.
https://github.com/EmidioStani/dcat-ap_validator/blob/master/rules/rule-108.rq
this one is a more complex case
Hugo: once a rule is in SPARQL it can be expressed in SHACL using the templates
Karen: yes using the templates SHACL can include any SPARQL query
Karen: does it take us anywhere?
Corey: not sure, other than there are non SHACL
Antoine: if there's something there that SHACL doesn't do then it's a bad sign because this project seemed to be in our core audience
ACTION: Karen to have a look at the rules and see whether they match our requirements and what SHACL can do.
Karen: Miika could be a good person for looking at SHACL. I will send him a note.
In future meetings, we should discuss how we see this connect to application profiles.
This group was tasked to further develop the DC concept of application profiles.
The original AP was developed with XML in mind.
Antoine: My simple answer (maybe too simple to be true): ... Wish I had the
time to do this.
ACTION: Antoine to articulate how SHACL relates to DC notion of application profiles.
Corey: Just six of us, and we are all busy - not best context for tackling a question of
broader significance. "Good thing to do" but not priority -- that means something.
Karen: What do you suggest the group do? Has it reached it's capacity?
Corey: I think it is an important group, but hard to find time to do anything beyond
that. Worry about trying to accomplish something more ambitious. That said, I
welcome a conversation about this on our next call.
Karen: Yes, next time. Adjourn? Next: in two weeks...
4. Coming conferences and collaborations
- has there been recent Hydra talks about RDF APs?
- Mariana's Call for Chapters - Book "Developing Metadata Shemas and Application Profiles"
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=DC-ARCHITECTURE;1377d958.1601
- DC 2016, TPDL 2016, other?
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=DC-ARCHITECTURE;d322e167.1601
5. AOB
===== On the backburner, until progress has been made on other work items (esp. requirements)
2015 Comparison of W3C and DCMI requirements:
https://docs.google.com/spreadsheets/d/1bCpQVyxI-N2Ca83umvQD8OKTdsDyG6Sz-E8Qo3v8ynM/
Hugo's email with suggestions: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=DC-ARCHITECTURE;41aa27ca.1505
ACTION: group to tackle by email:
- requirements taken from W3C which apparently don’t have a match in DC
- suggestions on the mapping from DC to W3C
Thread on W3C/DC requirement coverage, with Thomas doc
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=DC-ARCHITECTURE;539e5b2b.1512
SWIB discussion: check whether some items can be brought up in the agenda, e.g. update DSP
http://wiki.dublincore.org/index.php/RDF_Application_Profiles/SWIB2015_breakout
===== Pre-SWIB agenda points not merged in agenda [material for chairs / coming calls]
X. Polishing 2015 requirements
ACTION: Antoine to look at unclear requirements in Hugo's email
--ONGOING: only one requirement remains unclear: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=DC-ARCHITECTURE;8c92de00.1506
X Requirement database archival
ACTION: Antoine to create dumps and link them from the wiki pages - Valentine will contact Stuart.
We couldn't upload it to the wiki.
Valentine: Stuart is the one who could help
X. Planning for AP development
Look at Corey/Tom/Karen document and see how it fits into this in terms of defining the scope of an AP.
http://wiki.dublincore.org/index.php/RDF_Application_Profiles/blendingDoc
Do we need a connection with the next LDP?
ACTION: Corey to record a picture before the Minnesota discussion, on AP, 'de-referencing', 'graph definition'. See https://etherpad.wikimedia.org/p/dcmi-ap-13-08-2015
X Lars' new requirement, Discussed in previous call: https://etherpad.wikimedia.org/p/dcmi-ap-18-06-2015
ACTION: Antoine to sum up and email the discussion. [on the differences with other requirements] And put the case and requirement in our database.
X. Express requirements as RDF, mapping between DSP and our requirements
X. Coordination with BIBFRAME
Thomas' evaluation of BIBFRAME AP: http://wiki.dublincore.org/index.php/RDF_Application_Profiles/BibframeAnalysis
X. Coordination with Hydra, Fedora
Hydra Metadata Working Group: https://wiki.duraspace.org/display/hydra/Hydra+Metadata+Working+Group
Portland Common Data Model: https://wiki.duraspace.org/display/FF/Portland+Common+Data+Model
ACTION: Karen: post this link to W3C group - on hold until W3C group gets to the right point