DOIs from University of Manchester Library (DataCite)

[Stian Soiland-Reyes]

Hi,


We had a meeting today with Emma Thompson and her colleagues at the
University of Manchester Library, about Manchester assigning DOIs
(through DataCite) to data publications.

Attending from "our" side was myself, Stuart and Sean.


The University have finally signed the legal obligations etc., and the
service was officially live from yesterday. We will be forwarded the
contract so that we can understand the obligations - it will be a
question of how much of the obligation will remain with the Library
(e.g. make the resolver work) and how much with the individual project
(e.g. to provide a landing page for a workflow).


We are in particular talking about a couple of myGrid projects being
the "test bed" for assigning DOIs for artifacts such as:

* Models etc. at SEEK / FAIRDOM
* Workflows etc in myExperiment
* Research Objects in FARR Commons (HERC Chipset)
* $SimonHarpersThing (Sorry, I don't remember the details.. usability videoes?)


We also discussed items such as "How many?" (we think ~1000s to start
with) and metadata (we already have what is minimally required in each
of the systems, e.g. authors/title/date/type).

The contract is signed for initially 3 years, which means the Library
is allowed to mint DOIs during that period. It is likely to be renewed
if successful - but in anyway the assigned DOIs would keep working
beyond the contract.

There is a DataCite API, with single and bulk registration over JSON,
which Stuart has already played with using dummy account.

The Library will sort out how to create sub-delegations of the
official account (the sandbox DOIs are deleted after a week) - they
currently just have a single username/password and can do individual
registration per email requests - obviously we would want API access
so that we can add our "Publish with DOI" buttons.

(Do we want to assign DOIs to like every workflow version already
existing in myExperiment? Their metadata might not be quite intended
for that kind of "official publication" - the "Make DOI" button would
ask to verify that before freezing down that version).

It is kind of open to the Library what we want to have after the / in
the DOI - as long as it complies to the characters [A-Za-z0-9-.] (or
something strict along those lines)

10.1234/workflow15
10.1234/d123112
10.1234/1337
10.1234/myexperiment.workflow.15

We think that as the purpose of the DOI is to be a long-term
persistent ID, it should not convey too much information about the
implementation (e.g. it should not say "escholar" for escholar
datasets) - but should for the purpose of responsibility have a
delegation prefix that the Library can clearly associate with say SEEK
and the right contact person.

It might be a good idea to keep the DOI subdelegation to system
numeric - therefore it is not a big glaring problem if say
myExperiment becomes myWorkflow or SEEK changes name to FAIRDOM -- but
that after the sub-delegation prefix one can be slightly more semantic
(but not hierarchical).

The Library would maintain its own list of who is assigned which
numeric subdelegation - keeping it numeric means we also avoid fights
over prefixes (should myExperiment have preference for "workflow" or
"wf"? Who should get "data"?)


E.g. if we assign Manchester systems like:

11 myExperiment
12 SYSMO SEEK
13 FARR Commons
14 FAIRDOM

Then we can have these registrations:

10.1234/11.wf2867.v3 to http://www.myexperiment.org/workflows/2867/versions/2
10.1234/12.data.1088 to https://seek.sysmo-db.org/data_files/1088
10.1234/13.ro.fisher98 to http://www.farrcommons.org/ro/fisher98/

We see strength in having somewhat mappable the varying bit - e.g.
wf2867 for /workflows/2867/ as it means the DOI is predictable for a
system even before registering (which is good for self-citation).


What we agreed is that we will work on formalizing that suggestion and
get back to them with a list of which prefixes and registration
procedure we want to follow. Basically we are the beta testers of
this service within the University, and

We are also to look up best practice on minting the identifiers and
find other institutions that have done similar - not very different
from http://www.farrcommons.org/rules/identity.html I guess.


Stuart is to catch up with the DataCite APIs, and then we can try to
register a couple of selected objects pretty soon. We should be able
to reuse any DataCite code between SEEK and myExperiment.

It might be that this is a thing to try out at the myExperiment Tick
Fest - adding a "Make DOI" button that simply makes a new version (the
previous one is then frozen) and registers it as a DOI that we can
show prominently on the page.



--
Stian Soiland-Reyes, myGrid team
School of Computer Science
The University of Manchester
http://soiland-reyes.com/stian/work/ http://orcid.org/0000-0001-9842-9718

[Carole Goble]

All sounds promising. I bumped into Ian mulvany from eLife and he has been working on best practice for allocating duos to data. I'll ping him.

Carole

Sent from my iPhone by
Professor Carole Goble
The University of Manchester
UK

> --
> You received this message because you are subscribed to the Google Groups "myexperiment-developers" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to myexperiment-deve...@googlegroups.com.
> To post to this group, send an email to myexperimen...@googlegroups.com.
> Visit this group at http://groups.google.com/group/myexperiment-developers.
> For more options, visit https://groups.google.com/d/optout.