UCL LIBRARY SERVICES The evolution of a repository

UCL LIBRARY SERVICES The evolution of a repository: policy decisions at UCL Martin Moyle Digital Curation Manager, UCL Library Services e-mail: m. moyle@ucl. ac. uk SHERPA-LEAP Repository Workshop, Birkbeck, 28 March 2007

UCL LIBRARY SERVICES Background q UCL Eprints launched in March 2004 q Part of SHERPA-LEAP q Eprints 2. 3 platform (migrating to Digi. Tool this year? ) q Holds ~2500 records q Recently awarded a cautious mandate, initially for the collection of copies of RAE 2008 submissions and UCL research theses q Bottom-up development - plenty of trial and error q Most current policies can be extrapolated from http: //eprints. ucl. ac. uk/ - FAQ, Deposit Guide, etc q Here: the few policies that we're most often asked about

UCL LIBRARY SERVICES 1. What to collect? q Considered: q OA orthodoxy - refereed journal articles only q horror stories about preprints "graffiti" q need to cater for different disciplines q could we afford to turn content down? q Decision: any research output q they share a University; they can share a repository. . . q material has a “refereed / unrefereed” filter q structured to allow search/browse by Department/Research Group q includes conferences held at UCL and journals run from UCL q Uncontroversial

UCL LIBRARY SERVICES What we've collected. . . q q q q Journal articles (1600) Book sections (300) Conference papers (200) Working papers (160) Books (50) Theses (50) Discussion papers (50) Research reports (40) Technical reports (30) Book reviews (10) Patents (4) Literature reviews (3) Manuscript catalogues (2) Music scores (2)

UCL LIBRARY SERVICES 2. Mediated deposit vs author deposit q Started out with author "self-deposit" q Slow progress. . . q 44 papers added by end 2004 q 4 "real" depositors added 19 of these; rest added by Library team q UCL researchers produced ~7, 500 eligible papers in this period q Articles always publisher pdfs, so had to write or reject outright q bad news for relations with early adopters q heavy content bias towards conference papers and reports q Process of checking deposits very time-consuming q Evidence of abandoned work in author workspaces

UCL LIBRARY SERVICES

UCL LIBRARY SERVICES q Moved to fully-mediated service early 2005 q Good points: q Got things moving: added 1100+ papers in 2005 q Keeps things simple for academics q We were spending a long time checking records anyway qa bit more cutting and pasting doesn't hurt q Mediating brings us closer to the depositors q Issues: q Scalability q. UCL’s annual research output exceeds 10, 000 papers/year q. UCL Eprints has collected 2, 500 deposits in 3 years q. If author buy-in continues to grow, resources need to come too - or this policy needs to change q. Multiple authors – large STM presence at UCL. . .

UCL LIBRARY SERVICES

UCL LIBRARY SERVICES 3. Citation-only records q Initially insisted on full text q couldn't see the point of records without full text in an OA context q concerns about resources q After a few months, relaxed to allow citation-only deposits q Main drivers: q demand for completeness of coverage from senior academics q less alienating than outright rejection when copyright-protected content received q if publishers "go green", we'll be able to attach their files quickly

UCL LIBRARY SERVICES Citation-only records q Has proved to be a resource-intensive decision q publication back-lists come in regularly - some of them pretty big q. A. Computer Scientist – 106 papers; only 19 are citation-only q. A. N. Epidemiologist – 128 deposits; 109 are citation-only q Worth it to keep people happy q more likely to use us again q we make our priorities clear when we acknowledge backlists q helped by work on batch uploads from external databases

UCL LIBRARY SERVICES 4. University publications database q Has collected data since c. 1997 q all departments are obliged to make annual return q not complete, but high compliance q records often poor q q q Revamped for RAE around the time we were building UCL Eprints We had hoped to use RAE as a driver, but found we had competition Compromise. . . Agreed different purposes: dissemination / administration Records are harvested from us daily; include links to eprints q one-stop for depositors q incremental update - reduces annual burden q Has helped to embed UCL Eprints into research workflows q our RAE contribution = good data quality, increased research impact q Have resisted calls for total data alignment

UCL LIBRARY SERVICES 5. Deposit agreement q Carefully worded q Depositor has authority to deposit; no rights are infringed q We will distribute the work; we can translate the work q Depositor can request withdrawal; we can withdraw q Mediated deposit: authors no longer see the Agreement q Good practice, but. . . q Was it ever legal? q Would we really pass a legal challenge on to an author? q Is an author likely to complain, post-deposit, about distribution and format migration? These are cornerstones of our advocacy. q What are the risks? q Likely to become more relevant as 3 rd party services grow q A grey area. No changes planned at the moment.

UCL LIBRARY SERVICES 6. Other policies q Eligibility q Any UCL researcher q Will accept work produced elsewhere q Work stays if they leave q Formats q Take what we're given; convert to pdf where practicable qformats under review (use of pdf; preservation/access formats) qrecently introduced 'redirects' to subject repositories q No structured keywording q Data entry, not cataloguing q Full-text search is enabled and papers over the cracks q Most usage comes directly from Google and other search engines q May become more of an issue as 3 rd party services grow

UCL LIBRARY SERVICES 7. Overview of deposit process q Any research output, by any UCL researcher q Where publishers involved, we encourage author drafts at acceptance q early dissemination for authors q fewer copyright issues for us q loaded as 'In press' when publication is pending q Deposit by email - send us an attachment q no Deposit Agreement q We check the copyright q we write to publisher if position is unknown q citation-only if there are problems q Text files are migrated to pdf and uploaded q Records harvested next day by UCL publications database q Requirement on us to review “in press” records for completion q not just once - very inefficient

UCL LIBRARY SERVICES Conclusion q Pragmatic decision-making q Bottom-up approach allows evolution q Tuned to meet the needs of UCL researchers q Few barriers to participation. . . q. . . because the Library does all the work q Serious scalability questions q Works for now q Further developments will follow if and when we secure a full mandate