The scientific articles we write are uniquely identified by Digital Object Identifiers (DOI). Many people believe that we also need unique identifiers for the authors of those articles. Some of the arguments why author identifiers could be very helpful are found in this interview with Geoffrey Bilder, this Researcher Identification Primer, and this FriendFeed discussion. There is also an increasing number of scientific papers on the topic, including:
- Scientific publishing: Identity crisis (Nature)
- What’s in a name? (EMBO Reports)
- I Am Not a Scientist, I Am a Number (PLoS Computational Biology)
- Are you ready to become a number? (Science)
- Metadata for the metaconsortium (Nature Genetics)
Although most people agree that we need author identifiers for scientists, many details of how this should be implemented are not clear. I’ve listed some of the issues below. If possible, please take a few minutes and answer the questions for yourself in this poll.
1. What is the purpose of an author identifier for scientists?
- Unique identifier
- Author profile
- Authentication
- Other:
An author identifier should obviously uniquely identify an author. Some people like to add other functions, namely the ability to add a profile (e.g. a web page listing all papers of an author) and authentication. I think that the latter two functions can better be provided by a combination of unique identifier and some of today’s tools (see also question #10).
2. What is the best name for an author identifier for scientists?
- Author ID (Scopus)
- Researcher ID (Thomson Reuter)
- Contributor ID (CrossRef)
- Digital Author Identifier
- Other:
We need a name for author identifiers. Some of the names above are already associacted with an existing (or planned) service. Just pick a name and stick with it.
3. The author identifier system should be used for:
- Authors
- Reviewers
- Editors
- Database submitters
- Bloggers
- Commenters
- Other:
I believe that the author identifier system should be applied to all of the above. But it doesn’t have to be implemented for all these uses at the same time.
4. Who should pay for the author identifier system?
- Journal Publisher
- Blogger
- Database maintainer
- Author
- Other:
Those that gain the most from the author identifier system should pay for it. But because I think that quick adoption of the system is important, I wouldn’t make authors pay.
5. Should author identifiers be managed by a central organization?
- Yes
- No
I think yes, and for the reasons that Geoffrey Bilder explained in the interview with me. In my opinion a distributed system would create too many new problems.
6. Which organization should manage author identifiers?
- CrossRef
- U.S. National Library of Medicine
- Other:
The main reason that many people are reluctant to have author identifiers managed by a single institution is that they don’t want a single publisher or other commercial entity to control this system. But both CrossRef and the National Library of Medicine have a good track record for implementing publishing standards. I would prefer Crossref, as the National Library of Medicine is only concerned with a subset of papers, i.e. the biomedical literature.
7. Should an author identifier system for scientists be based on OpenID?
- Yes
- No
I would say no. OpenID is a distributed system, which is not desirable as discussed in questions #5 and #10. And OpenID also handles authentication, something which isn’t necessarily required (as discussed in question #1).
8. Should there be one or several author identifier systems?
- One
- Several
- Several with one top system
Similar to the DOI, I want only one unique author identifier. But this one unique identifier can then be linked to several other systems, e.g. an OpenID for authentication or a Nature Network, LinkedIn or Mendeley profile to list all your papers and other contributions.
I hope that enough people also answer these questions in the poll I posted.
Minor suggestion – perhaps one of the answers for the OpenID question should be: “what’s an openID?”
Cameron, I’ve added the option “what’s an OpenID” to question #7. And I will of course post the results of the poll. The poll closes after the first 100 responses, which hopefully is in a few days.
Brilliant, Martin: thanks for doing this. Certain of us are worrying about this very issue from the other side, as it were.
Right Richard, you now work for the other side, whatever that means…
It is his destiny.
I’ve given in to my—
no, no I’m not going there.
Hi Martin, interesting survey (just completed) question 2 is a bit loaded. You’re asking for the best name but then giving the actual systems – openid isn’t listed here. You might also be interested in Okkam . As for which organisation should managed identifiers, I think the author is ultimately has to be much more involved. For example, assigning DOIs to journal articles isn’t that personal, and doesn’t require authors to be invovled. However, assigning an equivalent id (e.g. DAI or contributor-id) to a person, is something authors need to be involved in I think… but I’m not sure exactly how yet…
Question 1: implies that Unique ID, Author profile, Authentication are all seperate things, but (I think) they are deeply intertwingled .
Question 2: “OpenID also handles authentication, something which isn’t necessarily required” why isn’t authentication required? You might want authors to authenticate publisher statements that they are indeed the authors of a given paper.
Duncan,
I think that most people want to give this author identifier a specific name to make the discussion about this topic easier. You could of course suggest OpenID.
You might want authors to authenticate publisher statements that they are indeed the authors of a given paper. You touch on the question whether author identifiers should also be assigned to papers already published, or only to newly submitted papers. In the first case we indeed have a problem of retroactively assigning authors to papers. The Science paper touches on this topic. Journal currently identify authors by email and signed consent forms, and I don’t see how this system can’t be used to also assign author identifiers. The journal could in addition also use a system like OpenID to authenticate, but that is a separate issue from author identifiers.
You are right that authors need control over their author identifier. But who should make the ultimate decision about the author identifiers assigned to a paper – the authors or the journal? I think it should be the journal.
Yes, I suppose the publishers can start the ball rolling, but the author needs to be involved to validate the claim. E.g publisher x asserts that person y is the author-of publication z.
But I don’t think the publisher should have the ultimate say, the person best placed (and probably most motivated) to validate such a claim is the author themselves.
Rod Page’s researcher-id page is a nice example of this. Although I don’t personally like researcherid, this is a good example of an author making a claim.
The only catch is, most authors can’t be bothered, but I think this is largely because they don’t realise how important the issue of digital identity of the web is…
…So the question becomes, how can we get authors to take their digital identity more seriously?
Duncan,
I have a slightly different perspective on this. I think the easiest way to start an author identifier system would be only for authors of journal papers and only for newly submitted papers. Author identifiers are used at the time of submission, and it is up to the journal that is handling the manuscript to assign an author identifier to every author.
Over time the author identifier will be used in other areas (for reviewers, blog authors, etc.). Software tools like those in use for Scopus or Web of Science will help in assigning author identifiers to papers already published, but here it is much more complicated to assign the correct author. The authors themselves would of course know best, but it is a time-consuming process and there is also the potential of fraud.
Martin, thanks again for doing this, I have participated in your poll.
Just a quick heads-up regarding contributor/author ID vs OpenID vs other ID systems: for those who didn’t already hear this (on FriendFeed or elsewhere), we are organizing a workshop to be held in Toronto May 13-14, titled Identifying Researchers on the Biomedical Web . See full details & registration form on the website (also the home of the Primer Martin refers to):
To name a few attendees, Geoff Bilder will come and present Crossref’s project, as well representatives from both Thomson-Reuters (ResearcherID) and Elsevier (Scopus Author Identifier) and also James Walker who will present digital identity + OpenID.
We want to not only discuss the author name problem, though it’s obviously a prominent topic, but the general concept of digital identity for researchers, how it could/should be used (community annotation, data access control, scientific blogging etc.) and mechanism for doing it (OpenID and friends, other tech.), and so on.
Gudmundur, thanks for mentioning the workshop, I hope that a lot of the material and discussions from the workshop will be available for those unable to attend. I will post the results from the poll when I have a few more votes (I’m currently at 42), maybe that’s interesting to the workshop.
The biggest question to me is whether or not to have OpenID as part of this author identifier system. Geoffrey Builder convinced me of the CrossRef ContributorID concept in the interview he gave to me a few weeks back.
Gudmundur, will someone from the Netherlands describe their Digital Author Identifier at the workshop?
Martin, making workshop materials, outcomes etc. available publicly is indeed our plan! I’ll definately post an announcement on our website and cross-link on FF when we have something. BTW for those not on FF: we also have RSS-feeds for new content and new comments , plus a newsletter you can sign up for to keep an eye on developments.
Regarding your other question: no, we won’t have anyone from that initiative at the workshop. I didn’t even know of this Dutch ID thing until a couple of days ago and therefore hadn’t contacted them about the workshop. Still, I reckon they would have heard about the workshop by now, e.g. via other people in Holland I’ve invited, and been in touch with us if they were interested. Will send them an E-mail anyway….
Hi Martin,
I had read here that researcher ID might have a EULA limiting its use outside Thomson. This fact stresses the importance of a solution that does not belong to a single “vendor”. Regarding OpenID, even though I recently noticed that the authentication need is really alive at least for publishers, I wonder whether the format chosen by OpenID is well designed to provide permanent ids. While permanency is also a matter of management and not only technology, the format name.webaddress as an id creates troubles when webaddress is lost, since name in itself it is not guaranteed to be unique. This is the general URL problem that standards like URN, LSID, Handle and ARK try to overcome. So even though the Single Sign On paradigm is still interesting, I guess OpenID can not as it is now be used as an identifier, while the protocol could be used on top of another id system. By the way, there is also an interesting development called FOAF & SSL for authentication.
I work for the OKKAM project (mentioned by Duncan), we are trying to define unique identifiers for authors in collaboration with already existing initiatives, contributor ID being the strongest at the moment IMHO considering the consortium backing up it.
Stefano, I completely agree with everything you say, especially about ResearcherID, OpenID and Contributor ID. I will post the results of the survey in the next couple of days and hope to provide some interesting material for discussion.
Stefano, there is movement within the OpenID community towards other forms of identifiers, like E-mail addresses. In fact, right now in the current 2.0 spec the ‘directed identity’ concept is used by Google, Yahoo and others, so users merely need to know that they have an account with Google and point the relying party there in order to authenticate. No need to know that your OpenID is a URL. There is a long, cryptic URL somewhere under the hood, but the user doesn’t need to know this. So, I do not think the URL format is necessarily a limitation in this regard.
But I agree with you (and have made this case elsewhere) that OpenID (and/or other SSO solution) could work on top of other systems, and users could have the option to link the various services (Contributor ID, identity ‘silos’ like Facebook and Myspace) together based on his/her identity.
I have some ramblings on this stuff here if you care to have a look.
The amount of discussion were are seeing in the last few months around author identifiers for me is a good sign that many people want this to succeed. We have many interesting initiatives and like many others I do have a ResearcherID, Scopus Author ID and OpenID.
What I hope to see in the next 6-12 months is a better integration of these services using a common researcher identifier provided by an independent institution, and a clarification of some of the issues discussed here and elsewhere (what is the role of OpenID, required vs. voluntary, who pays for this service, etc.).
I would bet that the CrossRef Contributor ID project has the best chances in succeeding. But it is important that we as researchers make our ideas and concerns known so that Contributor ID is not just simply built around the needs of scientific publishers.
I’ve posted the survey results.. There is some interesting stuff for discussion in it, e.g. the fact that quite a number of people don’t seem to know OpenID (the answer to question 7 suggested by Cameron). This is surprising to me because I assumed that the people answering the survey questions would be very much interested in this identity discussion and would know about OpenID, ResearcherID, etc.
Hi all. For those who are interested, we now have a provisional agenda for the IRBW2009 workshop. See the workshop page on our site.