PR for using ORCIDs to give credit

[GoogleCodeExporter]

Suzi suggested that we might do  "a little PR about the proposed use of ORCIDs 
to give credit to people for their contributions is needed. Ascribing credit 
for ontology contributions has been a pebble-in-the-shoe (boulder?) for at 
least a decade. It's fantastic that a solution is now being put into place."

I agree that this is a good suggestion for the outreach committee, but we will 
first need to finalize our policy on using ORCIDs.

Original issue reported on code.google.com by rlwalls2...@gmail.com on 26 Oct 2012 at 8:03

[GoogleCodeExporter]

As per last (20121024) outreach group meeting:
ORCID proposal - Alan
See 
https://groups.google.com/forum/?fromgroups=#!topic/obo-foundry-outreach-working
-group/wdcWWH8zwfE

Form at 
https://docs.google.com/spreadsheet/viewform?formkey=dEVGVVd6NHk5bFJrZ0QzbDJZVU5
OMUE6MQ#gid=0

AI: Alan will write draft and send to group.

Draft will ask for ORCIDs now, and will promise a design document that says how 
we will use it (initial idea: That is the URI for a person in attribution 
fields) in 1.5 - 2 months. That would be put out for comment, and then a final 
policy will be drafted based on feedback.

Original comment by mcour...@gmail.com on 26 Oct 2012 at 8:46

[GoogleCodeExporter]

From technical WG call 20121109

We should provide a template for people to add metadata for their ontologies, 
and they’ll create their IDs at this point if they don’t have one yet. No 
need to bug them specifically at this stage.

Original comment by mcour...@gmail.com on 9 Nov 2012 at 6:11

[GoogleCodeExporter]

Remember to make any recommendations backwards compatible with existing 
practice and usable in oboedit.

For GO we are likely to implement something such as:

 * An rdf version of http://www.geneontology.org/doc/GO.curator_dbxrefs
  - make use of standard vocabularies like dc, foaf, IAO
  - owl:sameAs for linking to ORCID, researcherid?
 * Some kind of bridge between existing axiom annotations and the URIs in this file
  - not straightforward in OWL-DL. May need to generate linking axioms, ie via sparql query

Original comment by cmung...@gmail.com on 2 Feb 2013 at 9:56

[GoogleCodeExporter]

regarding: "owl:sameAs for linking to ORCID, researcherid?"

We should avoid having more than one identifier. Thompson participates in ORCID 
and will provide services for mapping between researchid and ORCID. I don't 
think the GOC: URIs - if there were any - are worth preserving. So: Perhaps add 
a mapping from GOC: ids to either ORCIDs or instances of Organization whose 
members have ORCIDS.

Original comment by alanruttenberg@gmail.com on 4 Feb 2013 at 8:12

[GoogleCodeExporter]

Original comment by rlwalls2...@gmail.com on 16 Feb 2013 at 12:20

• Changed state: Started

[GoogleCodeExporter]

There are ~200 personal GOC IDs. So far I've mapped one of those to an ORCID 
(mine). It may take a while to do the other 199, especially as many of the 
people have moved on...

Original comment by cmung...@gmail.com on 16 Feb 2013 at 1:28

[GoogleCodeExporter]

Hi all,
I need to add ORCID IDs to an ontology we are working on. I still don't have a 
clear idea of what annotation properties should be used. Can someone advise? 
Even if it is an interim solution I would like to get something going.

Thanks,
Melissa

Original comment by haen...@ohsu.edu on 19 Apr 2013 at 4:06

[GoogleCodeExporter]

Right now we have  <dc:creator xml:lang="en">Carlo Torniai</dc:creator> for ex 
(taken from OBI file)

One option would be to create a new annotation property - maybe <orcid> - in 
IAO, and assert an annotation on the creator annotation. Issue is that it 
forces you to maintain the name Carlo and his associated ORCID.
Other option would be to not create anything, and just use directly the ORCID, 
which would end up as <dc:creator 
xml:lang="en">0000-0000-0000-0000</dc:creator>. Drawback is that it is not 
human friendly, advantage is that you don't duplicate info.

Ideal solution IMO: use <dc:creator 
xml:lang="en">0000-0000-0000-0000</dc:creator>, and have ORCID provide useful 
RDF at that URI http://orcid.org/0000-0000-0000-0000. When releasing the 
project you could pull in info from there if you want to add human names. 

Minor notes:
- I don't know how http://orcid.org/ is stable or not as base URI, so I would 
suggest that unless we know we just use the ORCID
- the ORCID search doesn't seem very efficient. I tried searching for Carlo, 
then myself and couldn't find us (I know at least myself have an ORCID, pretty 
sure Carlo does too)

Temporary solution: we create an RDF file with that info that can be pulled in 
at release time if desired. The ontology just uses <dc:creator 
xml:lang="en">0000-0000-0000-0000</dc:creator>

Original comment by mcour...@gmail.com on 19 Apr 2013 at 4:44

[GoogleCodeExporter]

Use the same ones we already use: definition source, term editor, dc:creator 
and doc:contributor at the ontology level. Make sure to include a rdfs:label 
for each of these so they display in a meaningful way. Call if you have any 
questions. There's no need for further properties, as far as I can tell. We had 
documented that these annotation properties would eventually be filled with 
instances, and certainly the dc: properties allow for both (I have seen both 
literals and instances used)

Original comment by alanruttenberg@gmail.com on 19 Apr 2013 at 5:12

[GoogleCodeExporter]

Don't you think it would make sense to have "extra" annotation such as 
rdfs:label etc living in a different file? Having all the information in one 
place would make it easier to update (and ensure there is no typo in name for 
ex), and possibly add more info - so for example Dr X is involved in 4 
projects, when he changes institution we update only in one place. Also gives 
room to add email or else, and have the change automatically propagated.

From the ontology developer point of view, you create dc:creator with the ORCID 
and just import a file called "researchers.owl" that we maintain.

Original comment by mcour...@gmail.com on 19 Apr 2013 at 5:55

[GoogleCodeExporter]

I like all of these ideas. One thing that we want to enable is easy search of 
an ontology by name or ORCID ID, so that it is useful to both human and 
machine. I think that would be solved by Melanie's modular method suggestion? 
we could have people request that their ORCID IDs be added to that file. Of 
course it would be helpful if there were user/ontology-specific configurations 
that could be enabled in Protege.

Whatever we decide, I would like to put up a wiki page with examples so that we 
have some easy-to-use documentation and help ensure consistent use. To discuss 
on next tech call? Perhaps we can even draft documentation at that time.

Original comment by haen...@ohsu.edu on 20 Apr 2013 at 3:25