**MACE Call 9-January-2012**

**Attending**
RL "Bob" Morgan, U. Washington (chair)
Ken Klingenstein, Internet2
Cliff Lynch, CNI
Amy Brand, Harvard/ORCID
Diego Lopez, Telefonica I+D
Renee Shuey, Penn State
Scott Cantor, The Ohio State U.
Rodney McDuff, U. Queensland
Chris Phillips, CANARIE
Jim Leous, Penn State
Michael Pelikan, Penn State
Chris Hubing, Penn State
Keith Hazelton, U. Wisconsin - Madison
Steven Carmody, Brown U.
David Wasley, Independent
Nick Roy, U. Iowa
Nate Klingenstein, Internet2
Ann West, Internet2
Steve Olshansky, Internet2 (scribe)

**Discussion**
Topic: Scholarly Identity
A number of complementary (or perhaps competitive) activities are happening in the area of enhanced identity for scholars and researchers. Projects such as VIVO, Catalyst, ORCID, SciENcv, and NSF Data Management Plans are moving forward and will have interesting requirements and opportunities for institutional and federated identity. See Ken's attached doc for more info and links.

Cliff began by noting that systems that maintain a faculty scholarly inventory (e.g. VIVO and Catalyst) are more heavily deployed in the UK and Australia than in the US. Systems like VIVO are intended to span institutions, v. the old model of keeping these sorts of systems local.

Key points:
- Data (e.g. NSF and NIH data mgmt plans, and from other funding agencies), bring issues involving identity. The ability to attach authorship to datasets so they can be cited, and so that the curators can be credited appropriately. Note DataCite, operating internationally.
http://datacite.org/

- What happens to data that is not entirely public (e.g. that containing PHI)? Access restrictions have particular properties: they are long lasting, they may span institutions, and may require both RBAC and user-level permissions. Roles are often not well articulated at participating institutions.

Attributing identity correctly for scholarly journals is a long-standing challenge. Various databases doing this are not sufficiently disambiguated. Connecting works published under different names for the same scholar is also a challenge (e.g. with/without middle initial, or name changes from marriage or divorce). It is becoming more common in tenure and promotion for article citation counts to be included.

Retrospective cleanup of the existing mess is something being tackled by GoogleScholar, Microsoft Academic Search, and others, but they are not doing strong identity proofing and each is an island (not propagating corrections between systems). Going forward, a unique author ID (e.g. from ORCID) would be a great benefit, but who assigns them and how they are issued to users are challenges to examine.

Many publishers are using journal submission mgmt systems, with their own IdM issues. It would be very useful to connect these to a federated identity system. The open source Open Journal Systems is apparently working on federation,
http://pkp.sfu.ca/?q=ojs

Most campus IdM systems don't accommodate names other than those in HR or SIS's rather than the names they publish under. Additional attributes reflecting published names were proposed as a useful contribution.

--
Amy reviewed ORCID as a registry of unique identifiers for contributors to scholarly works. Their phase one system is planned for release soon. Their focus is on knowledge discovery and assembling public profile metadata, not really security and privacy. They are interested in engaging with InCommon.

Q: What about ORCID as a service for campuses or publishers?
A: They are developing a membership business model, with fees based upon numbers of researchers at an institution. It is expected that researchers will find efficiencies in using the system that will attract them to use it.

Q: Relationship or overlap with VIVO? Missions overlapping? Do users need to choose one or the other?
A: Coordinating efforts with VIVO, but technical work on metadata schema and interoperability is yet to come. VIVO has commissioned ORCID to work on profile exchange, but the work is still early. These systems live outside the institution, and do not constitute a user's entire profile. VIVO talks to ORCID, but other systems do as well. They are very complementary.

Q: Where on campus to find those interested in this, and evangelize?
A: The libraries on campus will often be the locus of interest in these sorts of issues and systems, as will groups concerned with tenure and promotion practices. The office of sponsored research will be interested in current CVs for faculty, as will the university relations office which e.g. fields requests to provide experts for media interviews (SciENCV is in this arena).

Amy noted that expertise search capabilities may be of interest to different groups on campus, but appropriate indexing and taxonomy to enable this sort of search is a key element to include in these systems.

Q: Alignment of ORCID consortium membership with InCommon?
A: ORCID includes publishers, funding agencies, and universities, all of which are potential InCommon participants.

Q: Exemplar institutions to promote?
A: ORCID cited U. Chicago and Harvard, funded by NSF to document profile exchange.

JimL noted that PennState has experts.psu.edu for finding experts in response to media requests. PennState Emerging Technologies is looking at a VIVO project.

Amy noted that finding scholars without conflicts of interest due to co-publishing is an issue to be tackled in many contexts.

Q: When scholars move between institutions, how does ORCID handle this? Is identity linking a challenge?
A: Institutional affiliation is used for seeding the ORCID database, but the ORCID ID is not wedded to that original affiliation. Cliff noted that it is instructive to work an identity linking scenario in detail. When a user moves to a new institution, it is provided with an ORCID ID, and the system then retrieves the scholarly record and populates it into the new org system. This might be pushed or pulled, depending upon the circumstances.

Since students often are involved in grants and have published prior to becoming faculty, there is a clear need to engage them early in their careers. It may not be far off for university student applicants to provide ORCID IDs...

Q: How does a user prove an ORCID ID is in fact legitimately hers?
A: Working on that, currently username/password...

Q: StevenC noted that increasingly blogs are the platforms for initial ideas in academic debates. Does this imply value in assigning ORCID IDs to blog posts?
A: No reason not to, but best to avoid judgments about what are legitimate elements of scholarly output.

In ORCID, the researcher designates fields as public or private, or otherwise controls access, but privacy profiles haven't really been tackled yet per se. Cliff expressed interest in references relating to what is or isn't considered "public." Published works and grant awards are generally considered public, but what about things like addresses or other PII? Many institutions have worked on classifying data WRT privacy.

Q: Some national libraries are assigning IDs to authors?
A: Not really, mainly limited to book authors and not journal authors, but they do do disambiguation. These efforts have largely been disconnected from the work discussed today, but OCLC is represented on the ORCID board.

**
NEXT STEPS
- ORCID has regular open participant meetings, anyone interested is welcome.
- Communicating with the ORCID interim technical director would be productive.
- CNI is willing to serve a convening function for future conversations.
- Archival IdM problems relating to datasets would be a useful topic for further exploration.
- The Internet2 Spring Member Meeting in April may be a useful setting, if the right people are in attendance.
- International issues? Invitational conference organized by the Knowledge Exchange on author IDs in London mid-March. CNI and ORCID will be attending. REFEDS would be a good place to evangelize this.
- NSF data management plan discussion on a future MACE theme call.