Friday, 25 July 2014

The old in and out . . . .

As well as using the institutional repository to manage your official record of outputs for things like REF and PDR, you will often want to utilise other reference management formats, in citing your work while writing, or importing records from other systems. 

The repository can accommodate a large number of these, with a couple available at launch and more developed over time. Initial requirements for the repository identified DOI and PubMed as the key online sources and EndNote and BibTeX as the key bibliographic formats. DOI and PubMed are available on the Sandbox now. Work is proceeding well on EndNote and BibTeX, which require modifications to work with our unusual repository structure (c.f. Behind The Curtain)  

Let us know in the comments below if they are working for you and what additional sources / formats you would like to see available.

Wednesday, 23 July 2014

Behind the curtain . . .

How you manage recording your publications is changing . . . this blog post will cover some of the processes involved in ensuring your existing information moves across seamlessly to the new platform.

First, a little about the existing facility - the TULIP Manage Publications Function (TULIP-MPF). This has evolved organically out of the University's changing needs to record and report research outputs. It allows a lot of flexibility in how information is recorded. This has its good and bad points. On the plus side, the lack of constraints allows any information to be entered. On the negative side, the lack of constraints allows any information to be entered . . .

For example, in TULIP-MPF you can enter any author names as free text, formatted any way you want. This is great for getting your presentation just right in the other places you use that record BUT it makes it very difficult to link back other members of staff that may be on the publication. There are also innumerable ways to enter the same name of a journal, which can make tracking WHERE people publish difficult.

As part of the move to the new system, there has been a concerted effort to rationalise all this free form information and provide a bit more structure so that, in future, information will be more rigourously controlled. For example, in the new system, as you enter an author's name, suggestions from the staff list will pop up. When you start to type a publication's title, if there are similar titles in the system, they will appear, so that you can see if a colleague has already entered it, and link to that, instead of creating a duplicate.

A lot of time has been spent trying to identify information structures and to carry them across, to try and reduce duplication and make the new information intuitive and useful - both at a personal, and institutional level. One area where this has required a compromise is the way TULIP-MPF records the people involved in a publication. In TULIP-MPF there are two types of person record - those involved with administering the record, and those on the citation (authors, editors, etc). Unfortunately, the authors is a free text string - that is, one of those flexible pieces of information. This means that there's no way for the new system to know which of the authors on teh citation are members of staff, so initially the new repository will only maintain those links to administrative users - which may, or may not, include authors. This means that, in the first instance, those authors who have colleagues on papers will have to edit the citation to link them directly to their staff record.

Initially, the project promises to deliver information AT LEAST AS GOOD as that which currently exists in TULIP-MPF. Over time, individuals will curate their records and much of the duplication and low quality information will be replaced and removed.

Interested in what your records will look like in the new system? The Sandbox has the records from TULIP-MPF and the Research Repository - why not head over and take a look, then let us know what you think in the comments below . . .

Friday, 18 July 2014

Climb every mountain . . . . a brief overview of the ongoing technical work delivering the new Repository

As we start the final transition from the TULIP manage publications to we are going to publish a number of technical articles explaining some of the processes and issues involved. This first one is a general overview and introduction.

The project is split into a number of phases. At the start of the upcoming academic year, phase one will deliver. This is a replacement of the functionality of the manage publications function in TULIP (TULIP-MPF), moving from an ad hoc internally developed system to one of the institutional standards. Subsequent phases will develop the features above and beyond those already available.

Phase one has focused on delivering a simple process to move all of the existing records out of TULIP-MPF (and the pilot Liverpool Research Archive). A great deal of effort has been focussed on structuring the information to make it more reliable. For example, in TULIP-MPF, the author names in a citation are free text. This makes it really hard to link authors and publications together, especially if a colleague enters the data. In the new system all personal data is directly linked to the staff records for UoL employees and will make suggestions based on records already existing for other contributors.

Autocomplete suggestions will search existing records for titles of publications already in the repository, so that the same output is not entered twice and journal names will also feature autocomplete suggestions.

The system will make it easier to enter already existing article records, initially supporting import from DOI and PubMed. Other bibliographic formats (EndNote, BibTeX, etc) will be available for import and export.

Upcoming articles will focus in a bit more depth on some of the above. If you have something in particular you would like to hear more about, why not leave us a comment?