This month we have introduced a new metadata format: KBART. Adding the KBART export format to both DOAB and the OAPEN Library has been requested by the library community and offers another option for exporting our metadata and seamlessly connecting to the metadata and to open access books. The KBART export format is now available alongside the other metadata feeds on the OAPEN and DOAB websites.
Selective Export
In addition to the KBART format we have introduced a ‘selective export’ functionality for the Directory of Open Access Books (DOAB) & the OAPEN Library. This enables users of both platforms to export books and chapters in a flexible way: first by selecting the relevant titles and secondly by choosing the export format.
The selective export functionality makes it easier for specialist libraries, subject repositories and frequent users to only retrieve specified subsets of open access books according to their preferences.Please be aware that this export function is optimised for sets up to 500 titles.
How to use the Selective Export
Step 1: Search for books
Step 2: If needed, you can refine the export even further
Step 3: Select the export format
Step 4: The metadata becomes directly available
If you have questions about these new functionalities or wish to share feedback, please contact Ronald Snijder (r.snijder[@]oapen.org).
You are probably familiar with USB cables, the type of standard cable that allows me to connect a printer and a phone to my laptop, using the same port. However, the cable I use for my phone does not fit my printer. If I were to use one cable, I would either have to buy a new printer or use some kind of connector.
This is similar to ONIX: an XML standard designed to exchange metadata about books between all the different actors in the book supply chain. Here, the standard also comes in different sizes. For instance, an author might be coded as <b037>Snijder, Ronald</b037> or as <PersonNameInverted>Snijder, Ronald</PersonNameInverted>. Moreover, the current version of ONIX is 3, but publishers who have invested quite a lot in version 2.1 might be reluctant to give that up. These differences in standards and approaches can result in a complex situation for actors in the ebook supply chain; more details in this draft report by the Book Industry Group.
So, at OAPEN we need to make sure that all those different flavours of ONIX fit our import system and therefore we built a connector. In fact, we built two: one that converts different types of ONIX XML to Excel, and one that converts Excel to ‘our’ ONIX standard.
Why two conversions? Because we need to do a manual check on the metadata we received: does it contain titles that are already part of the OAPEN Library collection? Is there – still – metadata missing? Is there metadata that needs to be amended? When the metadata has been checked and updated, we convert the resulting Excel file into ONIX XML and import it into the OAPEN Library.
If you are interested in trying this out yourself, please mail me at r.snijder[@]oapen.org. But if you want to wait for a more user-friendly and better developed tool, you should head over to the COPIM project. In this project, development is on the way of Thoth, a tool to manage open access book metadata: https://copim.pubpub.org/pub/open-metadata-thoth/release/1.
Connecting different things – whether appliances or systems – requires work and is never easy. But we are trying to create a better fitting solution.
It all started with a preprint. Back in February 2020, I came across the article “Increasing Visibility of Open Access Materials in a Library Catalog: Case Study at a Large Academic Research Library”. Here, Jeff Edmunds and Ana Enriquez describe their work to highlight existing and add more open access titles to their library catalogue at Penn State University Libraries. Simply put, they investigated ways to make it easier for library patrons to discover open access books, by making them more visible in their library catalogue. Library catalogues ingest metadata based on the MARC standard, and the authors discussed recent developments regarding MARC and open access.
Their timing was perfect. The OAPEN Library was to be migrated to the DSpace 6 platform, and this was the right moment to revise our metadata offerings. Thus, I contacted Jeff Edmunds and asked how we could improve our MARC metadata feeds. This was the start of an involved conversation about the general quality of our metadata records and the presence of OA markers to make clear that the title described is open access.
Based on the recommendations, we updated our “MARC exports” in several ways. Firstly, each MARC record now contains both a link to directly download the book – or chapter – and a link to the landing page in the OAPEN Library. In this way, readers have the option to first assess the book in the OAPEN Library or they can directly get the title. Another improvement was to ‘move’ the license URL to MARC field 540, which helps libraries to automatically process in what way the books can be used. Furthermore, to make sure that the cataloguing software automatically flags the OAPEN Library content as open access, we added specific markup to the 506 and 856$7 fields.
The result
The result of all this? In August, Penn State University Libraries added close to 10,000 new open access records to their catalogue – based on the contents of the OAPEN Library. The import was based on this MARCXML feed, and converted to MARC21 using MarcEdit. To allow for the character limit of the identification field used by the cataloguing software, the HANDLE identifier used by OAPEN – for instance https://library.oapen.org/handle/20.500.12657/25287 – was shortened to “oapen25287”.
At Penn State University Libraries, this is how it looks in “The CAT”, searching on “oapen library”:
Penn State University Libraries are transitioning from their old ‘classic’ catalogue to a new layer built using Blacklight. The plan is to add an Open Access facet or search limit in the new catalogue, which will allow users to find all OA content as a group. In the updated catalogue, the results look like this:
When we look at a single book record, the open license is clearly visible. Also, there are multiple links – in this case a link to the landing page in the OAPEN Library and a link to directly download the PDF version and a separate link that opens the EPUB version of this book:
The catalogue can be searched online, using this link.
Starting September, Penn State Libraries will update their “OAPEN collection” monthly.
Future wish list
While we are quite happy with the results, there are still some things that can be improved. To start with a technicality, the MARC Leader field is lacking a few characters at the end:
“=LDR 02817naaaa 00349uu” should be “=LDR 02817naaaa 00349uu 4500”. More important is the planned conversion of the Directory of Open Access Books (DOAB) to DSpace. The new DOAB platform will have the same export features as the OAPEN Library, including the MARC and MARC XML feeds.
Ideally, the descriptions in the OAPEN Library and DOAB would have name headings in an authorised form and WorldCat subject headings. These features will remain on our future wish list, at least for now.
Add the OAPEN Library collection to your catalogue?
Our cooperation with Penn State University Libraries was successful, especially because of the guidance we received regarding our MARC metadata output. We hope that other libraries will follow their example and add the OAPEN Library collection to their catalogue. If you have any questions, please contact me via r.snijder[@]oapen.org, or drop me a line on Twitter @ronaldsnijder.
The OAPEN Library has become a highly connected resource for full text open access books and chapters: used by the Directory of Open Access Books (DOAB); libraries; data aggregators and search engines – for instance Google Scholar.
Consequently, the quality of the metadata is important to us and when it came to our attention that not all authors and editors were listed in the correct order, we started looking for a solution. This is not a trivial matter: at this moment the OAPEN Library contains over 12,000 books and chapters.
To complicate matters, the metadata – collected for a decade – describing these publications come from a variety of sources. A manual check was not an option, so we searched for a resource that would enable us to check the author and editors listed.
We decided to use the freely accessible CrossRef API. OAPEN is a member of CrossRef – and has deposited hundreds of DOIs – but more importantly, the CrossRef metadata database lists over 1.4 million books.
First, we made a selection of over 2,200 books and chapters with more than one author or editor. Using the DOI connected to these publications, we retrieved the ORCID metadata. Then we checked the metadata found against our own records, resulting in a correction of over 1,000 title descriptions.
Every now and then we will perform this and other metadata checks, to make sure that everything remains… in order.