UoL Library Blog

Develop, debate, innovate.

Repositories and CRIS WRN event article

Posted by gazjjohnson on 23 August, 2010

Nick Sheppard, Leeds Met University (aka MrNick on twitter) has written a good article in the most recent Ariadne about the Welsh Repository Network/JISC workshop back in May looking at the interaction between CRISes* and repository systems.  As I was unable to get to this event due to prior commitments, it was good to have a chance to catch up on the discussions.

I was interested to note that a CRIS (Current Research Information Systems) can go by many names – given that the UoL Research Office often refer to them as RIMS – Research Information Management Systems.  They’re not alone as many universities seem to have renamed them as RMAS or ERA and the like.  But at their heart they are systems that not only gather in research publication data (and much more), but actively link to other systems – chief among them from my perspective interlinking with a repository.

The question “Is an IR a subset of a CRIS?” posed by one speaker (Simon Kerridge, ARMA) is an interesting one.  Having seen a number of recent CRIS vendor demos, it is one that is clearly approached in different ways by different organisations.  Some very much see the IR as a satellite system, fed largely (but not entirely) by the CRIS.  For others it is more of a subsumed system – with a visible front end peeking out, but the rest of the body absorbed by the greater whole.  I must confess so long as the workflows for such issues as rights verification and data management are still handled by the elite repository administration team I don’t have an especial problem either way.  However, if a CRIS/Repository union means that a repo is just a reflection of the CRIS data set, locked down without the additional resources embodied and ingested by the IR over and above the REF related items; well then I’m a little more uneasy.

The talk from St Andrews’ Data architect Anna Clements (which came with some interesting but not readily comprehensible diagrams) brought up the CERIF standard.  Interesting that St Andrews has been pursuing links to their repository for far longer than many other institutions, which has demonstrated the advantages of working closely together with research support personnel (something I’ve benefited from here at Leicester in the past two years and can heartily concur).

Meanwhile William Nixon and Valerie McCutchean of Glasgow gave a very useful overview of the integration of the repository with a CRIS.  I was able to plot from my own experiences whereabouts we are in this process here at Leicester.  They raised a valuable point about author authorities – something that has long concerned me as an issue to which I don’t have a ready solution.  In some regards I’m hoping the CRIS implementation here will allow us to tackle and resolve this at that point – given that unique IDing of authors is something that is key for bibliometrics and REF returns alike.  I notice William doesn’t appear to have offered a solution though in his talk, which is perhaps a slight concern for me.  I wonder how difficult it is going to be to match an author of a non-REF item that routes into the repository from beyond the CRIS with the institutional verfiied author list.  And what about external additional authors?  I suspect this is going to be a major issue for me and my team to resiolve and one that I’d welcome external insight on.

Finally my old friend Jackie Knowles talked about the pitfalls of implementation – most of which I am, thankfully, already well aware.  I think we definiely need more of these warts and all case study examples though; as at the end of the day those of us working at the sharp end of repository/CRIS interlinking will need to know how to work around so many of them.

It sounds like this was an excellent day (and perhaps in serious need for near future repeating!) and a definite must read artilce for anyone about to establish, or already working towards, a CRIS/Repository interlink.

8 Responses to “Repositories and CRIS WRN event article”

  1. Nick said

    Hi Gaz

    Regarding unique IDing of authors – the JISC Names project is promising I think –


    • Cheers Nick – I’ve been aware of the Names project for a while – but still not quite got my head around “what they can do for me”! Wondering if there’s a neat button perhaps they can add to my repo that will solve all my ills…

      • Hi Gaz

        Plugins for the repository software are not available yet, but are definitely something that we are planning to develop in a future phase of Names. One of the things Names needs in the meantime is better information about UK researchers, so if there is information about Leicester researchers that you are able to share with us, to improve our (currently rather sketchy) data, then we would be delighted to have it!


        • Nick said

          Hi Amanda

          Regarding information about institutional researchers, what would be the best way for us to liaise with the Names project?

          I’m perhaps being naive(!) but if I could get a BIG LIST of names from HR would that be useful – what sort of info (email, department)? What sort format – just a long list of names in a spreadsheet?


          • Nick – even just a simple Excel list with full names will be useful, I’m sure. Departmental information would make it more useful (might allow matching with existing data). Up to now we haven’t got any affiliation information in the system (as the Zetoc data doesn’t hold that), so if we can use Leeds Met’s staff as a guinea pig test-bed, that would be great. We’ve hit Data Protection concerns at this point with other institutions, though, so you might need to investigate that…


  2. William Nixon said

    Further details about our author work can be found on our Enlighten blog

    We use our local Glasgow Unique ID to ID our local records and then use that ID to link the member of staff to the publication. The name shown on our author Browse list is the name used in our HR system [and which appears on payslips, the Staff A to Z etc] the link enables us to deal with name variations, e.g. maiden/married as well as citation preferences such as initials rather than full names.

    See: Author browse list

    When records are manually added, the depositor should identify the Glasgow author(s) We can use EPrints autocomplete to show author names and e-mail addresses. Library staff will also review author listings. For some areas like Physics, we are bulk importing the data and the Glasgow authors have already been identified so that they are automatically assigned the appropriate GUID, this is vital in examples like this:

    A recently added feature now provides links from Glasgow authors to their publications, in the above example, it is now apparent for instance that Bussey and Robson are Glasgow authors. Their names link to their own author publication lists in the IR.

    There is a crosswalk table which also maps the EPrints user ID, which is the public number/link to the GUID and staff number ID’s which are not public to enable other University systems to surface publication data for an author, in other parts of our website.

    We don’t do any further matching on external authors.

    For REF, we envisage the IR providing the platform for REF selection – but this would be a [confidential] subset of their complete output.

    Our CRIS doesn’t act as an intermediary or a proxy for delivering content to our IR – no publications data is held in our IR. It’s a complementary system which we pull our project and funding information from to populate our Funder View

    • Cheers for this William, always appreciate your sage advice! That sounds like that works well in practice, so I think it assuages some of my concerns – although how we get to such a beutiful point might be a bit more of a struggle. I guess I’m worying though still about those odd items that come to us direct (non-REF bits) and how much work that’ll be to match Prof XTC with Dr X.Tc on the paper for my team. Something that will only really come out in practice though as we move into an implementation and operational phase of CRIS/Repo activity.

      • William Nixon said

        Taking a step back there are two key issues here which really only in the last year we have now “cracked” [or at least are on the way to] as part of our JISC funded Enrich Project.

        The first is ensuring that you have an authoritative list of staff data which to draw on. We were able to achieve this by working with our IT Services department who plugged our University data vault [which is fed by HR data] into the Repository. This gave us more than 12,500 user accounts (which included Honorary and Associate Staff) – with their GUID, Staff Number, Department and e-mail address. This system is working very well.

        The University has recently restructured and all staff received new e-mail address and college/school data – now that we are plugged into the Identity Vault, all of that data was updated in the IR, so user records show the new e-mail account.

        The 2nd element and in some cases the more complex which we have found is mapping, especially with backfilled data, the person to the publication. We have identified records which are currently “orphaned” i.e. there are no links and are doing some work to match these on a college/school basis. Having the Honorific has been incredibly valuable and if you can use an autocomplete feature which will show the college/school staff belong to then also increases the matching certaintity.

        In many cases we have more publications for a member of staff than may be currently listed under their browse view because older records are not yet linked – so it’s not that we don’t have the publication, it’s just there is no link.

        The flip side of this is that we now have the ability/opportunity for staff to readily see the list of publications attached to their name and to identify any which may require updating or in some cases which may be wrongly assigned to them.

        As new records come into the IR, we have found that the bulk of them have at least one author with a GUID, in many cases more than one so that these are much more straightforward and when they go into the live archive they automatically bind to the author’s publications listing.

        Best wishes for your ongoing implementation.

Leave a Reply to gazjjohnson Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: