Skip to content

Instantly share code, notes, and snippets.

@stain
Last active January 8, 2016 21:52
Show Gist options
  • Star 2 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save stain/abaf89cf40df2823dde1 to your computer and use it in GitHub Desktop.
Save stain/abaf89cf40df2823dde1 to your computer and use it in GitHub Desktop.
2015-12-02 ORCID RDF telcon

ORCID RDF telcon

When: 2015-12-02 11:03

Who: Stian Soiland-Reyes (Univ of Manchester), Phil Archer (W3C), Tom Demeranville (Orcid)

Agenda

New actions

  • ACTION: Tom to review the 303 pull request and put into QA
  • ACTION: Stian to simplify Jena code usage so it's easier to modify vocabularies
  • ACTION: Tom to Modify URIs used for account and profile document
  • ACTION: Tom to Keep Phil/Stian/Ivan updated on RDF progress (due end of 2015)

Skype transcript

See transcript.html or transcript.txt below. (HTML generated using http://philarcher.org/tools/skypescribe/ )

  • Update to Jena 3.0.0, improve Jena usage, avoid ontology loading
    • Provide JSON-LD output
    • JSON-LD context?
  • Add schema.org properties/type
  • Fix / vs # - add <> and #account to distingish foaf:OnlineAccount and foaf:ProfileDocument
  • Change 307 redirect to 303 See Other
  • Investigate how to represent bibliographical profile, FaBIO/ SPAR ontologieS?
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en-GB">
<head>
<title> ORCID-RDF chat </title>
<link rel="stylesheet" type="text/css" href="http://philarcher.org/css/skypescribe.css" />
</head><body><h1> ORCID-RDF chat </h1><h2 id="present">Present</h2><p class="people">System, Stian Soiland-Reyes, Phil Archer, Tom Demeranville</p><h2 id="regrets">Regrets</h2><p class="people">None received</p><h2 id="chair">Chair</h2><p class="people"> Stian-Soiland-Reyes </p><h2 id="topics">Agenda Items</h2><p><a href="https://gist.github.com/stain/abaf89cf40df2823dde1">Agenda</a><ol></ol><table><tr><td colspan="2" class="info">Meeting began at 11:01:22</td></tr><tr><td class="speaker">scribe</td><td class="comment"> Stian-Soiland-Reyes </td></tr><tr><td class="speaker"></td><td class="comment"> </td></tr><tr><td class="speaker">hello</td><td class="comment">guys </td></tr><tr><td class="speaker"></td><td class="comment"> </td></tr><tr><td class="speaker"></td><td class="comment">shall we start? </td></tr><tr><td class="speaker">ok</td><td class="comment">we&#39;ll wait a few more minutes for Phil </td></tr><tr><td class="speaker">Phil Archer</td><td class="comment">OK I&#39;m all set </td></tr><tr><td class="speaker"></td><td class="comment"> </td></tr><tr><td class="speaker">Stian</td><td class="comment"> Phil mainly pointed out some issues on URIs. Phil: URIs identify different things. Person is not bytes. Person is not updated last week! So that&#39;s why we need to know what we are talking about. The actual person, an account, or the cat they own, etc. All different URIs. What Sian did originally was to make the differentiation the person and account by adding a / to the end of the URL. The problem of that it is easily confused - generally the browser does this. I stood up in front of academic people and said this was terrible! And then I had a look at it.. and see.. no, it&#39;s not the same! Just one character difference. I was a Research Data Alliance and talked to Laura, and we talked about this. So I said I would offer some suggestions, which ORCID would either ignore or take on board. That was a long time ago! <a href="http://www.w3.org/2015/03/orcid-semantics">http://www.w3.org/2015/03/orcid-semantics</a> </td></tr><tr><td class="speaker">Tom</td><td class="comment"> Talking to this with the dev team.. it&#39;s easy enough to add such changes to the implementations </td></tr><tr><td class="speaker"></td><td class="comment">&hellip; hoping we would get it right. But we&#39;re OK with doing whatever you recommend. </td></tr><tr><td class="speaker">Phil</td><td class="comment"> Problem is that any change must be.. Do we have any sense of existing users of the data? Tom: Stian: Like we have been using ORCIDs in our RDF, but without programmatically accessing the info (but relying on the ability to do so) Tom: We have some other version of the API.. and RDF can get a bit outside the versioning here. Phil: Also this is a bit secret.. no indication anywhere that this existed. Stian: Like a button somewhere Phil: Like I suggest how to improve navigation. Tom: Redirection is already part of the API.. but it is still marked experimental and not part of the documentation. So improving this description is one of the tasks we have there. </td></tr><tr><td class="speaker">Tom</td><td class="comment"> As it works differently from the other APIs it is a bit different from the REST APIs </td></tr><tr><td class="speaker"></td><td class="comment">&hellip; But it will be become part of the API </td></tr><tr><td class="speaker"></td><td class="comment">&hellip; Adding extensions is a bit tricky.. like we have more than one .json representation. But we could do like .ttl </td></tr><tr><td class="speaker">Stian</td><td class="comment"> If it&#39;s .ttl or /experimental-rdf or so is not as important.. Tom: Will end up on one of our API endpoints. Could you clarify something? </td></tr><tr><td class="speaker">Tom Demeranville</td><td class="comment"><a href="http://orcid.org/0000-0000-0000-0000 should 303 to <a href="http://pub.orcid.org/v1.2/0000-0000-0000-0000">http://pub.orcid.org/v1.2/0000-0000-0000-0000</a> ">http://orcid.org/0000-0000-0000-0000 should 303 to <a href="http://pub.orcid.org/v1.2/0000-0000-0000-0000">http://pub.orcid.org/v1.2/0000-0000-0000-0000</a> </a></td></tr><tr><td class="speaker"></td><td class="comment"> So the first URI is the person, the second is the profile document </td></tr><tr><td class="speaker"></td><td class="comment">this solves the / issue </td></tr><tr><td class="speaker">Phil</td><td class="comment"> One worry here with the second URL is the version number.. </td></tr><tr><td class="speaker">Tom</td><td class="comment"> Our API is versioned.. </td></tr><tr><td class="speaker"></td><td class="comment">&hellip; it&#39;s not critical to get rid of the version number, the big thing is the redirect </td></tr><tr><td class="speaker">Tom Demeranville</td><td class="comment"> curl -i -H &#39;Accept: application/rdf+xml&#39; <a href="https://orcid.org/0000-0003-4654-1403">https://orcid.org/0000-0003-4654-1403</a> </td></tr><tr><td class="speaker"></td><td class="comment">Serif? </td></tr><tr><td class="speaker">Phil Archer</td><td class="comment"> CERIF </td></tr><tr><td class="speaker"></td><td class="comment">different URI structure needed for CERIF as it&#39;s a separate API </td></tr><tr><td class="speaker"></td><td class="comment">Only person info is returned in the ORCID </td></tr><tr><td class="speaker"></td><td class="comment">what about publications </td></tr><tr><td class="speaker"></td><td class="comment">Stian can&#39;t do: foaf:publications as then the person is also a page (foaf:publication range is a foaf:Document) but perhaps foaf:publications with corresponding id=&#34;publications&#34; in the HTML </td></tr><tr><td class="speaker">Phil Archer</td><td class="comment"><a href="http://xmlns.com/foaf/spec/#term_publications ">http://xmlns.com/foaf/spec/#term_publications </a></td></tr><tr><td class="speaker">Stian</td><td class="comment"> IDeally also publications could be in RDF,, but more work would be needed in choosing ontologies etc. here </td></tr><tr><td class="speaker">Phil</td><td class="comment"> My concern is to push you in the right direction </td></tr><tr><td class="speaker"></td><td class="comment">At W3C in our spec-writing tool we are trying to make it easy to include people&#39;s ORCID </td></tr><tr><td class="speaker">Stian</td><td class="comment"> with PROV statements of the spec? </td></tr><tr><td class="speaker">Phil</td><td class="comment"> One of the requirements we found internally.. consensus was that requirements of linking to a profile, it must return RDF (among other things) </td></tr><tr><td class="speaker">Stian Soiland-Reyes</td><td class="comment"><a href="http://fantasai.inkedblade.net/style/design/w3c-restyle/2016/sample ">http://fantasai.inkedblade.net/style/design/w3c-restyle/2016/sample </a></td></tr><tr><td class="speaker"></td><td class="comment">? </td></tr><tr><td class="speaker">Phil Archer</td><td class="comment"> Example of W3C usage of ORCID <a href="http://w3c.github.io/dpub-pwp/">http://w3c.github.io/dpub-pwp/</a> </td></tr><tr><td class="speaker">Phil</td><td class="comment"> I&#39;m talking from a theoretical point of view, while you are on the practical side. So it has to work for ORCID&#39;s users </td></tr><tr><td class="speaker"></td><td class="comment">&hellip; Just get the basics right.. then the community might ask for more :) </td></tr><tr><td class="speaker">Tom Demeranville</td><td class="comment">Summary Change to 303 redirects Change foaf:publications to {html url}#activities Change personalProfileDocument to {api url} </td></tr><tr><td class="speaker">Tom Demeranville</td><td class="comment"> Add documentation </td></tr><tr><td class="speaker"></td><td class="comment">303 redirect fix: <a href="https://github.com/ORCID/ORCID-Source/pull/2472/files">https://github.com/ORCID/ORCID-Source/pull/2472/files</a> </td></tr><tr><td class="speaker">Phil</td><td class="comment"> ... crossref etc.. should be semantically aligned </td></tr><tr><td class="speaker"></td><td class="comment">&hellip; Also talking to Ms Yan on CERIF in another projects.. Very small world! :) </td></tr><tr><td class="speaker">Tom</td><td class="comment"> Get some basic Dublin Core </td></tr><tr><td class="speaker">Stian</td><td class="comment"> Will try to improve the Jena usage in the code so it&#39;s easier to expand vocabulary use - no need to use the ontology support in Jena really. </td></tr><tr><td class="speaker"></td><td class="comment">&hellip; You also mentioned provenance? </td></tr><tr><td class="speaker">Tom</td><td class="comment"> Yes, about activities.. DOI relationships </td></tr><tr><td class="speaker"></td><td class="comment">e.g. a record can be imported several ways </td></tr><tr><td class="speaker"></td><td class="comment">tricky for us. Early on we didn&#39;t distinction between how it is added. </td></tr><tr><td class="speaker"></td><td class="comment">if datacite add something to datacite.. is that because you did it, or because a datacite bot did it </td></tr><tr><td class="speaker">Phil</td><td class="comment"> sit on UK Research Data Forum. Nick Right from NEwcastle. </td></tr><tr><td class="speaker"></td><td class="comment">&hellip; Lots of what we talk about here.. orcid, crossref, datacite.. getting the semantics usable, is relevant to that group </td></tr><tr><td class="speaker">Timelines</td><td class="comment">Tom is trying to get it into testing by end of year. </td></tr><tr><td class="speaker"></td><td class="comment">as changes are quite minor </td></tr><tr><td class="speaker">Phil</td><td class="comment"> Hoping a small delta would have some impact :) </td></tr><tr><td class="speaker"></td><td class="comment">Ivan Herman is on some advisory boards for ORCID </td></tr><tr><td class="speaker"></td><td class="comment">both Phil and Ivan are happy to poke around and review </td></tr><tr><td class="speaker">Phil Archer</td><td class="comment"><a href="http://philarcher.org/tools/skypescribe/ ">http://philarcher.org/tools/skypescribe/ </a></td></tr><tr><td class="speaker">Stian Soiland-Reyes</td><td class="comment"><a href="http://s11.no/s/ ">http://s11.no/s/ </a></td></tr><tr><td class="speaker">System</td><td class="comment">skype ?chat&blob=2sQ3HkRVO04dDjG8jS2ayP6REi2Yj_iRp04CMjRdNI4tqBhxFqGp7ZTgePxA4yBT_0wDrYgfXi-LOMl9 </td></tr><tr><td class="speaker">Stian Soiland-Reyes</td><td class="comment"><a href="http://s11.no/s/?chat&blob=2sQ3HkRVO04dDjG8jS2ayP6REi2Yj_iRp04CMjRdNI4tqBhxFqGp7ZTgePxA4yBT_0wDrYgfXi-LOMl9 ">http://s11.no/s/?chat&blob=2sQ3HkRVO04dDjG8jS2ayP6REi2Yj_iRp04CMjRdNI4tqBhxFqGp7ZTgePxA4yBT_0wDrYgfXi-LOMl9 </a></td></tr><tr><td class="speaker"></td><td class="comment"> </td></tr><tr><td></td><td class="action"><span>ACTION:</span> Tom to review the 303 pull request and put into QA </td></tr><tr><td></td><td class="action"><span>ACTION:</span> Stian to simplify Jena code usage so it&#39;s easier to modify vocabularies </td></tr><tr><td></td><td class="action"><span>ACTION:</span> Tom to Modify URIs used for account and profile document </td></tr><tr><td></td><td class="action"><span>ACTION:</span> Tom to Keep Phil/Stian/Ivan updated on RDF progress (due end of 2015) </td></tr><tr><td colspan="2" class="info">Meeting ended at 11:42:11</td></tr></table><h2 id="actions">Actions</h2><dl><dt> Tom</dt><dd>to review the 303 pull request and put into QA </dd><dt> Stian</dt><dd>to simplify Jena code usage so it&#39;s easier to modify vocabularies </dd><dt> Tom</dt><dd>to Modify URIs used for account and profile document </dd><dt> Tom</dt><dd>to Keep Phil/Stian/Ivan updated on RDF progress (due end of 2015) </dd></dl></body></html>
2015-12-02 ORCID-RDF chat
Present:Stian Soiland-Reyes, Phil Archer, Tom Demeranville
Chair: Stian-Soiland-Reyes
Agenda Items:
Agenda
Meeting began at 11:01:22
scribe Stian-Soiland-Reyes
hello guys
shall we start?
ok we'll wait a few more minutes for Phil
Phil Archer OK I'm all set
Stian Phil mainly pointed out some issues on URIs. Phil: URIs identify different things. Person is not bytes. Person is not updated last week! So that's why we need to know what we are talking about. The actual person, an account, or the cat they own, etc. All different URIs. What Sian did originally was to make the differentiation the person and account by adding a / to the end of the URL. The problem of that it is easily confused - generally the browser does this. I stood up in front of academic people and said this was terrible! And then I had a look at it.. and see.. no, it's not the same! Just one character difference. I was a Research Data Alliance and talked to Laura, and we talked about this. So I said I would offer some suggestions, which ORCID would either ignore or take on board. That was a long time ago! http://www.w3.org/2015/03/orcid-semantics
Tom Talking to this with the dev team.. it's easy enough to add such changes to the implementations
… hoping we would get it right. But we're OK with doing whatever you recommend.
Phil Problem is that any change must be.. Do we have any sense of existing users of the data? Tom: Stian: Like we have been using ORCIDs in our RDF, but without programmatically accessing the info (but relying on the ability to do so) Tom: We have some other version of the API.. and RDF can get a bit outside the versioning here. Phil: Also this is a bit secret.. no indication anywhere that this existed. Stian: Like a button somewhere Phil: Like I suggest how to improve navigation. Tom: Redirection is already part of the API.. but it is still marked experimental and not part of the documentation. So improving this description is one of the tasks we have there.
Tom As it works differently from the other APIs it is a bit different from the REST APIs
… But it will be become part of the API
… Adding extensions is a bit tricky.. like we have more than one .json representation. But we could do like .ttl
Stian If it's .ttl or /experimental-rdf or so is not as important.. Tom: Will end up on one of our API endpoints. Could you clarify something?
Tom Demeranville http://pub.orcid.org/v1.2/0000-0000-0000-0000 ">http://orcid.org/0000-0000-0000-0000 should 303 to http://pub.orcid.org/v1.2/0000-0000-0000-0000
So the first URI is the person, the second is the profile document
this solves the / issue
Phil One worry here with the second URL is the version number..
Tom Our API is versioned..
… it's not critical to get rid of the version number, the big thing is the redirect
Tom Demeranville curl -i -H 'Accept: application/rdf+xml' https://orcid.org/0000-0003-4654-1403
Serif?
Phil Archer CERIF
different URI structure needed for CERIF as it's a separate API
Only person info is returned in the ORCID
what about publications
Stian can't do: foaf:publications as then the person is also a page (foaf:publication range is a foaf:Document) but perhaps foaf:publications with corresponding id="publications" in the HTML
Phil Archer http://xmlns.com/foaf/spec/#term_publications
Stian IDeally also publications could be in RDF,, but more work would be needed in choosing ontologies etc. here
Phil My concern is to push you in the right direction
At W3C in our spec-writing tool we are trying to make it easy to include people's ORCID
Stian with PROV statements of the spec?
Phil One of the requirements we found internally.. consensus was that requirements of linking to a profile, it must return RDF (among other things)
Stian Soiland-Reyes http://fantasai.inkedblade.net/style/design/w3c-restyle/2016/sample
?
Phil Archer Example of W3C usage of ORCID http://w3c.github.io/dpub-pwp/
Phil I'm talking from a theoretical point of view, while you are on the practical side. So it has to work for ORCID's users
… Just get the basics right.. then the community might ask for more :)
Tom Demeranville Summary Change to 303 redirects Change foaf:publications to {html url}#activities Change personalProfileDocument to {api url}
Tom Demeranville Add documentation
303 redirect fix: https://github.com/ORCID/ORCID-Source/pull/2472/files
Phil ... crossref etc.. should be semantically aligned
… Also talking to Ms Yan on CERIF in another projects.. Very small world! :)
Tom Get some basic Dublin Core
Stian Will try to improve the Jena usage in the code so it's easier to expand vocabulary use - no need to use the ontology support in Jena really.
… You also mentioned provenance?
Tom Yes, about activities.. DOI relationships
e.g. a record can be imported several ways
tricky for us. Early on we didn't distinction between how it is added.
if datacite add something to datacite.. is that because you did it, or because a datacite bot did it
Phil sit on UK Research Data Forum. Nick Right from NEwcastle.
… Lots of what we talk about here.. orcid, crossref, datacite.. getting the semantics usable, is relevant to that group
Timelines Tom is trying to get it into testing by end of year.
as changes are quite minor
Phil Hoping a small delta would have some impact :)
Ivan Herman is on some advisory boards for ORCID
both Phil and Ivan are happy to poke around and review
Phil Archer http://philarcher.org/tools/skypescribe/
Stian Soiland-Reyes http://s11.no/s/
System skype ?chat&blob=2sQ3HkRVO04dDjG8jS2ayP6REi2Yj_iRp04CMjRdNI4tqBhxFqGp7ZTgePxA4yBT_0wDrYgfXi-LOMl9
Stian Soiland-Reyes http://s11.no/s/?chat&blob=2sQ3HkRVO04dDjG8jS2ayP6REi2Yj_iRp04CMjRdNI4tqBhxFqGp7ZTgePxA4yBT_0wDrYgfXi-LOMl9
ACTION: Tom to review the 303 pull request and put into QA
ACTION: Stian to simplify Jena code usage so it's easier to modify vocabularies
ACTION: Tom to Modify URIs used for account and profile document
ACTION: Tom to Keep Phil/Stian/Ivan updated on RDF progress (due end of 2015)
Meeting ended at 11:42:11
Actions
Tom
to review the 303 pull request and put into QA
Stian
to simplify Jena code usage so it's easier to modify vocabularies
Tom
to Modify URIs used for account and profile document
Tom
to Keep Phil/Stian/Ivan updated on RDF progress (due end of 2015)
@essepuntato
Copy link

About the investigation on "how to represent bibliographical profile, FaBIO/ SPAR ontologies" introduced in https://gist.github.com/stain/abaf89cf40df2823dde1#file-suggested-orcid-rdf-updates-md, I would be happy to help if needed (I'm one of the co-author of SPAR)!

Please consider that I'm already involved in something similar with DataCite, and maybe it would be good to join forces in that.

@stain
Copy link
Author

stain commented Dec 2, 2015

My suggested RDF clean-up as a pull request: ORCID/ORCID-Source#2475

@stain
Copy link
Author

stain commented Dec 2, 2015

Brilliant, @essepuntato! Agree that SPAR could work well, but we are a bit limited in our rich the descriptions are in ORCIDs, so I guess some selected properties would work well.

You might be interested in these classes:

@stain
Copy link
Author

stain commented Dec 2, 2015

@essepuntato, you might want to start from my rdf-uri-fixes branch from pull request ORCID/ORCID-Source#2475 - as it lets you drop in vocabularies in orcid-api-common/src/main/vocabs to let Jena generate constants.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment