Repositories and OpenID

This is a short post... it should be longer, but I don't have time to flesh out my thoughts properly and I wanted to at least get something down on paper (so to speak).

Pete and I recently met up with Traugott from UKOLN for a quick drink.  Just catching up really, but in our conversation we got round to talking about name authority in the context of scholarly publishing and institutional repositories.  In a rash moment I asked, "Why don't we just use OpenIDs as author identifiers in institutional repositories?".

I'm not sure that Traugott was very impressed... possibly for good reason!  He was particularly concerned about legacy issues for example.  But it seems to me that the idea shouldn't be dismissed totally out of hand.  I've felt for some time now that any centralised approach to name authority is pretty much doomed to failure for all sorts of reasons that I won't go into here.  I've had at the back of my mind that one might be able to build a distributed solution using LDAP, i.e. based on the LDAP servers maintained by institutions.  But it seems to me that using OpenIDs has some significant advantages:

• Firstly, academics don't consider institutional repositories, or even academic services, to be their sole focus of attention.  They are equally interested in all sorts of Web-based services alongside the stuff delivered by their institution and the more formal academic services that they get access to by virtue of being members of an institution.  Therefore, as I've argued before, any access and identity management solution needs to work across the whole range of services that academics are interested in if it is to be compelling.
• Secondly, academics have an online life before and after their academic career and they move between institutions during it, so anything that is tied too closely to a particular institution is problematic.

OpenID is nice because it is so distributed, open and flexible.  Academics wouldn't be forced to use the identity given them by their current institution, though they might choose to for various reasons.  But an external identity offered by a third-party would be able to migrate across institutions seamlessly and so has some significant advantages.

Note that academics would not be forced to use a single OpenID for everything they do if they didn't want to.  For example, they might choose, for privacy reasons, to use different OpenIDs for academic and non-academic services.  But clearly they would have the option of using a single OpenID if they wanted to, and doing so would carry with it significant advatages in terms of single sign-on.  Overall, the flexibility of OpenID, and its decentralised approach, would theoretically leave the end-user far more in control of how their online identity(ies) were used.

OK, so it's still relatively early days in the OpenID story, but let's say that OpenID becomes the normal way in which your average academic identifies him/herself to the range of external Web 2.0 tools they are interested in using (their blog and so on).  Wouldn't it make sense for them to also use an OpenID as an author identifier?  The same OpenID that they use to log into their institutional repository.

Wouldn't that be cool?

Comments

[Phil Wilson]

For what it's worth, I've been considering enabling OpenID accounts for every user at the University of Bath for a few weeks now, so http://www.bath.ac.uk/person/819130/id would be my OpenID URL. I'll see if I can make any progress on this in the next few days.

Feel free to mail me if you have any further ideas or suggestions, I'd be more than happy to hear them!

[Seb Schmoller]

And not just academics. Why not learners as well?

Seb

[Jessica Lindholm]

Except for the fact that your posting brings back nice memories from several pub visits with you, Pete and Traugott at different occasions in Bath, a comment I read a while ago at http://iwantmyopenid.org came to mind.

"I'd love to see OpenID make inroads in higher ed (a place that could really use such a system). The DSpace and Fedora projects are persistent object repositories used in higher education, cultural institutions, and industry. [---] Since higher ed assumes it is going to throw you (and your credentials) out every few years, these notions fundamentally clash. OpenID's flexibility would allow people persistent, trusted access to their objects stored in Dspace and Fedora, without institutions needing to keep credentials around longer than they need/want to (Going to graduate? Just verify your OpenID to your repository profile.). [---] Since the problem OpenID would be solving so relevant, DSpace and Fedora might make good candidates." http://iwantmyopenid.org/bounty#comment-4

Can't figure out who made that comment, but it seems to illustrate the same thoughts.

As I am working with a DSpace-with-LDAP-implementation at Malmo University nowadays, persistency for authors (not just employees) seems to be a thing to look more into.

[Roddy MacLeod]

Is there not some way that any new name authority initiatives builds on what is already available - e.g. Community of Scholars?

"Community of Scholars provides direct access to more than 1 million active researchers working in over 200 disciplines and 8 countries. Searches in Community of Scholars provide unparalleled exposure to the community currently studying a topic of interest. Exploring this living research community expands awareness and access to cross-disciplinary research perspectives, publications, research methodologies, expertise, and collaborative opportunity. With Community of Scholars, your faculty and students are directly connected to the community of scholarly research.

Scholars within Community of Scholars are also linked to their publications in select 3rd party databases and databases from CSA, facilitating direct links to author background, contact information, and linked publication histories to provide enhanced document context and novel community based pathways to relevant literature."

[AndyP]

Phil, thanks for the info... I'd be interested to hear about your experiences as you roll this out.

Seb, yes... definately. The exact same arguments can be made about learners as I make here about 'academics'.

Jessica, thanks for the pointer.

Roddy, yes... things like Community of Scholars are important in having already done some of the leg work to solve the legacy problem, i.e. matching up author names on existing published scholarly material. My post is intended to be more forward looking - how do we exploit the Internet in such a way that we are not still in the position we are in now (essentially having to take educated guesses about when two authors are the same person) in another 10 years time. It's also not clear to me how open systems like Community of Scholars are? But, yes, in theory one could imagine that such services could become OpenID providers in the future.

[Bill Fitzgerald]

The piece that should also be addressed is what the OpenID url points to -- given that an OpenID can also be delegated to an external url (as explained here: http://tinyurl.com/y69g4s )

So, the authenticating authority becomes less relevant. Taking this a step further, a person can use their OpenID identifier to point to a location that actually means something; ie, a person's OpenID should point to a person's homepage, or profile, or electronic portfolio, or some combination thereof -- the beauty of OpenID is that it is actually a distributed identity system that lets a user control their identity.

Combine this with a system where learners and academics can easily move their resources/networks between systems and you're really on to something.

[Phil Wilson]

Or at least they could point to a URL as well as providing the OpenId delegation used the same methodology for pointing to FOAF, RSS, etc. etc. - not to mention the attribute exchange possible in OpenId 2.0 (https://verify.sxip.com/demorp/)