e-theses Harvesting

Background

In liaison with the National Library of Wales and the EThOS project the Welsh Repository Network is aiming to establish a full service electronic thesis harvesting system taking theses from the university repositories and ingesting them into both the National Library of Wales archival system and a central, public facing open access repository. In particular, this activity requires significant consultation across the partners as to the model to be used, the costs involved and getting buy- in from the institutions concerned.

Current work on this theme includes a survey of e-theses activity across the sector and work on the technical component of the harvesting service along with discussions with partner projects and institutions about the nature and scope of the service. The key points for discussion raised so far are as follows:

• The proposed harvesting model within Wales was agreed as possible but both the National Library of Wales and EThOS.
• Draft workflows have been documented.
• Theses collected in electronic format at point of submission have no cost implications for institutions (no digitization required) and will automatically feed into the harvesting process. It is the older paper copies of theses where digitisation is requested via EThOS that will need decisions taking at a local level.
• The proposed bid for a retrospective theses digitisation project will go forward via WHELF to the Welsh Assembly Government.
• Bibliographic control will be a big driver for the retro digitisation project with NLW providing high quality, preservation standard records.
• The NLW involvement will need fully mapping out and agreeing via some form of official paperwork.
• The paperwork trail both within institutions, and across the wider network as a whole, will need to reflect the model we adopt. This may take significant time to agree and put in place.
• UKETD metadata will need to be provided by the repositories at each institution. There will probably be some discussions needed with how this overlaps with cataloguing print theses.
• Risk approach for each institution will need to be decided (copyright checking and liability) - some consistency may need to be pursued across partners?

Metadata

In order to provide the essential interoperability between our local repositories, the National Library of Wales and EThOS, institutions will be required to store metadata for their e-theses within the repository using the UKETD_DC metadata standard. Further information about metadata requirements for e-theses are available via the EThOS Toolkit

In order to facilitate understanding of metadata requirements of e-theses the WRN team have set up test metadata records, one of these is a best practice record for an e-theses that is as complete as possible and contains all possible fields. A copy of this test record can be found in the document store. The test records will be used both for training purposes and for testing the movement of data around the system once the technical components have been installed on your server.

While the test records illuminates the possibilities and richest data possible for an individual theses it is noted that in practical terms much of this extra data might not be available, or institutions may not have time or resource to enhance records to this level. Consequently, we shall be looking for institutions to meet at least the minimum requirements of the UKETD_DC standard, this is just the following six fields.

• Title
• Author
• Awarding Institution
• Type
• Qualification Level
• Thesis Date

Further training and information about metadata requirements will be provided by the WRN time in due course.

E-theses Q&A

The following are a series of useful questions which have been raised by WRN partners during the course of discussions to date.

When an institution sends off an older printed thesis to be digitise by the EThOS service it is intended that the resulting digitised theses can be sent back to the institution concerned for addition into the repository. What are the implications that need to be considered when an institution starts receiving back these e-theses?

EThOS take a fairly laid back approach to risk management when it comes to copyright and permissions involved with digitised theses and much of the onus lies with individual institutions to conduct copyright checking of content prior to sending off theses to be digitised.

While most institutions can offer some level of checking on the copyright material contained within the text, they are not normally contacting the student concerned to get them to sign up to all the various rights we need them to sign to subsequently distribute the thesis via the repository and harvest it into the NLW e-theses portal we are building. This is in contrast with currently produced theses where the rights required can be obtained from the student at point of submission.

One suggestion that has been offered was that the originating institution should not subsequently download the full text of these retrospectively digitised and should avoid putting them into their own repository in full text, rather they should create a metadata only record which points to the full text on the EThOS site – therefore passing liability back to EThOS. This liability would only be waivered to a degree, as ultimately the institution took the decision to deem it suitable for digitisation. But, if it is obviously just the EThOS version that is primarily on the web, then it is only them that needs contacting in order to effect a rapid ‘take-down’ thus mitigating risk more effectively.

There are further issues to be considered with these older theses being digitised, in particular relating to the harvesting cycle, and the WRN team are currently in discussions with the NLW and EThOS regarding these. More information will follow as soon as we have it.

Does the student declaration form need to be returned to the library in paper form or could it be submitted electronically? Do you know if it would be permissible for us to keep a copy of a scanned form and signature on record instead?

Electronic signatures are a tricky area and typically they are not legally binding without investment in a whole raft of technical solutions to ensure the security of the signatures and identities of those signing. So, in practical terms, yes institutions should still have student declarations submitted in printed format. When it comes to then scanning of such the forms for keeping long term records the WRN team have checked with copyright experts and would advise to err on the side of caution advising that it would be best practice to keep the original printed signatures. However, there might be local practices which differ from this and we would suggest you check with your own legal experts if in doubt.

Would theses be stored in the repositories indefinitely or is there a time limit?

This would be a decision taken locally as part of the repository policy framework. The WRN team can offer training sessions on repository policies, contact us if you are interested in scheduling one of these. When it comes to the harvested copies of e-theses which the NLW will be gathering from around Wales, our understanding is that the NLW will be looking at preserving them in the long term, although the mechanics of this are still under discussion as we set up the service.