Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.



Section


Column
width50%


Panel
borderColorred
borderWidth3
borderStylesolid
titleBackground Documents

RDF Schema/OWL Binding for the DDI Richard Cyganiak

RDF Binding and Specification and Validation - from Copenhagen Sprint, November 2015

Survey of Past Work - from Dagstuhl Week One, October 2017

RDF Decisions and Open Issues - Written by Arofan Gregory in Dagstuhl Week One, October 2017


Warning
iconfalse

RDFS/OWL Meeting Minutes

Expand
title2018-03-20
Attending: Achim Wackerow, Eric Prud'hommeaux, Oliver Hopt, Jay Greenfield, Kelly Chatain
Oliver - demonstrates the XML tranformation as it currently exists. Everything used for transformation is in bitbucket DDI-Views_production/transform/  
The word doc gives an overview of the current views (Oliver needs to update it a little bit)
Need to discuss the gap between the documents on specific views (text docs) and in the XMI, which classes should be included in a specific view. The classes bring in more classes because of associations.
Some classes that don't appear in the View lists, but are in EVERY view. How is this documented?
Need documented rules for the transformation programs.
We need a feasible path for the prototype - Achim needs to discuss with TC or MT? Probably MT/Doc group. Kelly will bring up on next call. 
Jay - May be too late for the prototype. 
Oliver to send links to all of the documents that were discussed today. Is the algorithm used for the stylesheets documented? Oliver will update by the end of this week.
Eric demos FHIR (190 classes that interrelate)
Jay - when flattened, we lose the visual. To keep it intelligible, decide what to put back inline. DDI 4 not near a good documentation solution. 
Eric - Taking classes with single references can become nested. 
Clear understanding of applied rules for XML first, then conversation with MT/Documentation - how to make life easier with Views, how to communicatie with prototype (while the views are quite complex).
Jay - More discussion about the algorithm needed. 


Expand
title2018-03-01

Will have first OWL representation ready for review by March 8th. Will take a look at that and then pass on to TC after next week's call.

Achim to work on rules based on the Dagstuhl documents.

Eric to write down rules of what he has done thus far. 





Expand
title2018-02-01

Feb. 1, 2018

Attending: Kelly Chatain, Eric P., Achim

Timeline and how to support the iterative process.

  • Generation of the vocabulary is the most important.
  • Can we use the class documentation?
  • Eric testing parsers for various languages. Should transformation work on current drupal?
    EA flavor XMI should be used.
  • Targeting the obsolete EA right now. Doing the work that makes it easy to switch to canonical.

Iterative process:

  • Achim on vacation Feb. 12-16th.
  • Eric questions: abstract/interface - reserved names or problem with tool?
  • Input: Are property names effectively globally unique? Eric can produce a list of properties that have the same names.

Action: Achim to write RDF Invite

Action: Create JIRA issue for modellers about global unique property names.

Action: Strategy template on wiki for RDFS (Kelly) 

Double check the rules review with Eric.


Expand
title2018-01-26

January 26, 2018

Attending: Kelly Chatain, Eric Prud'hommeaux, Achim Wackerow, Jared Lyle

Logistics and Work Organization:

--Payment and support - Eric to submit the invoice to Jared on a monthly basis. Details are in the contract.

--Possible weekly call with following attendees: Achim, Eric, Kelly, Deirdre, Darren, Dan Smith, Guillaume, someone from modelling. Wendy or Dan Smith from the technical committee, specifically in regards to the production framework. 

--Set up an email list or group function with an archive function as a reference.

      Action: Kelly will look into JIRA options, eric will provide an example of a W3c mailing list?

-- Wiki: Kelly will set up a team page on the wiki, JIRA issue tracker, we need a bitbucket site for RDF work

Achim's list of issues outstanding in RDF work: (link to document)

--Reasonable chance that JSON LD will do what we need without too much effort.

--OWL binding/syntax – We are talking about the validation of the OWL model, then the examples.

--Validation in RDF, to make it falsifiable, we need to keep track of the things that can be disjoint which takes more effort. Syntactic validation is possible for the prototype (classname conflicts, etc.)

--Prioritizing issues for prototype and then parking the others.

--Documentation for validation needs to be in ‘english’ not computereese. In reStructuredText? Similar to markdown

--Re: Xslt to transform xml to rdf – would the xslt be a deliverable (user-friendly and tested)? Yes. All tools developed for our purposes should be part of the production framework, which will be published. The prod framework will be used by others to produce other bindings, or flavors of a binding. Should we use Javascript or scala or something else that is easier to read. Requires a different toolchain, but what is the most popular one?

      Action: Should be discussed in the RDF group and document discussion.

      Action: Any issues that we did not think about before, please document and specify the amount of effort needed for the prototype.

     Action: Who will debug and fix the xslt? – Oliver Hopt

--How can generic rules be used by a transforming tool so that the tool can be changed easily in the future. Can we do that currently for the prototype? Should we move away from xslt now or in the future? The cost is higher to carry it forward, so we should move (Eric) Transforming instance data to instance data is what more people do. Transforming the model is a smaller community. Eric is quick at the javascript.

     Action: Send Eric the latest version of the xml that Drupal spits out, the corresponding OWL, then the rules document (cygniak and Copenhagen)

--Are predicates unique across the system or within a particular class? We need to dive in a little more on this issue. If we had a converter, we could see when we were reusing predicates, then we would know how to make decision about migrating a few predicates or to say they effectively have the class implicitly prepended onto their name.

Example:

In one class is a person structure, but in another class is a string.

      Action: Kelly to confirm Eric JIRA permissions, bitbucket, etc. Will send emails to everyone.


Next call : Feb. 1st at 5pm CET with Achim, Eric,




Attachments


Column


Panel
borderColorgreen
borderWidth5
borderStylesolid
titleCURRENT WORK

Canonical XMI with documentation of how it is produced

Eric's development tool for exploring the model - Used in web browsers, Chrome best.  Drag the second dot (in the row of 4 dots) with the mouse to the URL form field. An URL should show up. After waiting for the end of processing the file, it is possible to explore the model in the browser.

protégé - recommended for exploring the DDI ontology 

Placeholder for current transformation rules

Placeholder for tracking diffs in canonical xmi

Placeholder for latest OWL/XML and ShEx representation



Panel
borderColorred
titleColorblack
borderWidth3
titleBGColorlightred
borderStylesolid
titleLinks to more information


Panel
borderColorGreen
borderWidth3
borderStylesolid

RDFS/OWL Strategy for TC Review - Linked


Info

This team meets weekly on Thursdays at 17:00 CET and 11:00 EST (daylight savings)

The next meeting is Thursday, April 12May 3, 2018.

Agenda: 

  1. TBD

Please join my meeting from your computer, tablet or smartphone. 
https://global.gotomeeting.com/join/775756413 

You can also dial in using your phone. 
United States: +1 (224) 501-3316 
France: +33 170 950 590 
Germany: +49 692 5736 7300 

Access Code: 775-756-413 

First GoToMeeting? Let's do a quick system check:https://link.gotomeeting.com/system-check