Refeds


Subject Re: The other side of academic identity
From Alex Reid <alex.reid@xxxxxxxxxx>
Date Tue, 29 Jun 2010 13:49:18 +0800

At 11:05 PM 28/06/2010, Alex Reid wrote:

At 09:25 PM 24/06/2010, Diego R. Lopez wrote:

I have been pointed to this site: http://www.orcid.org/

The consistent, unique identifiers that they claim for researcher citation
would benefit, and probably contribute as well, federated identity.
...
We have introduced auEduPersonSharedToken into our Australian Federation to serve just this purpose (I wasn't aware of ORCID at the time - early 2008).
...
I'll provide a bit on Use Cases when I can reach the AAF website (seems unreachable at present...).

Here are some pointers to Use Cases:

How researchers are using AAF: http://www.aaf.edu.au/index.php/supporting-researchers/how-researchers-are-using-the-federation/

Basic Use Cases at time of creation of auEPST:
---------------------
A core feature of Shibboleth federations is the existence of a privacy-preserving identifier – eduPersonTargettedID (ePTID). Each user has a different ePTID for each Service that they visit. Once an ePTID value is created for a user for a particular Service, this identifier is persistent for this Service over time. The key feature of this approach is that where appropriate and desirable, the user’s privacy can be maintained while at the same time allowing for an ongoing access to the Service that is specific to this user.

However, ePTID may not be best suited to certain requirements. The essence of these (based on ARCS and ANDS documents) are:

1. Identification across multiple services
  1a Identification across web-based services
  1b Identification across Grid services

2. Identification over time (persistent), including when a user changes institution (ie, portable)[1]

Both of these may be required at the same time (identification across multiple services when a user changes institutions).

[1] The ANDS requirement for an ID associated with a metadata record is here interpreted as a unique user account linked to the creation (and future editing) of a metadata record. This is opposed to an assumption that a human identifier is being stored IN the metadata itself (as this concept has both longevity and privacy problems associated with it).
-------------------

Cheers, Alex.

-------------------------------------------------------
T Alex Reid
Advisor, eResearch & Middleware
AARNet (Australia's NREN)
and
Honorary Professorial Fellow
School of Computer Science & Software Engineering
The University of Western Australia.
home address  71A Raymond Street, Yokine, WA, 6060
ph  +61 8 9345 0440
mobile  +61 40 888 5515
email  alex.reid@xxxxxxxxxx