Image Courtesy of AJC under a CC BY-SA 2.0 license.
This is part three of my Linked Data Series. You can find the previous posts in my author feed. I’ve decided to spice things up a bit and let you hear from some library professionals who are actually implementing and discussing Linked Data in their libraries. These interviews were conducted via email and are transcripts of the actual interviews, with very minor editorial revisions. This first interview is with Allison Jai O’Dell.
Allison Jai O’Dell is Metadata Librarian and Associate University Librarian at the University of Florida, George A. Smathers Libraries. She is on the editorial teams of the RBMS Controlled Vocabularies and the ARLIS/NA Artists’ Books Thesaurus – and is working to publish both as enriched, five-star linked datasets. Learn more about her from her website.
Can you give a brief description of TemaTres?
TemaTres is a free, open-source content management system for knowledge organization systems (KOS) – such as library thesauri, taxonomies, ontologies, glossaries, and controlled vocabulary lists.
Can you list some key features of TemaTres?
TemaTres runs on a Web-server, and requires only PHP, MySQL, HTML, and CSS. TemaTres is quick to install, and easy to customize. (Gosh, I sound like a salesperson! But it really is simple.)
TemaTres is a cloud-based solution for multiple parties to build and access a KOS. Out-of-the-box, it provides a back-end administration and editing interface, as well as a front-end user interface for searching and browsing the KOS. Back-end users can have varying privileges to add, edit, or suggest concepts – which is great for collaborative projects.
TemaTres makes it easy to publish Linked Data. Concepts are assigned URIs, and the data is available in SKOS and JSON-LD formats (in addition to other formats, such as Dublin Core and MADS). Relationships can be established not only within a KOS (where reciprocal relationships are automatically inferred), but also to external Web resources. That is, TemaTres makes it easy to publish five-star Linked Data.
How have you used TemaTres in your institution? Can you give an example?
I have used TemaTres on several thesaurus projects to streamline collaborative workflows and publish (linked) data. For example, at the University of Florida, George A. Smathers Libraries, we are using TemaTres to develop, publish, access, and apply local controlled vocabularies and ontologies. I am particularly excited to collaborate with Suzan Alteri, curator of the Baldwin Library of Historical Children’s Literature, to develop an ontology of paratextual features. Because our special collections are so unique, we find need to extend the concepts available in major library thesauri. With SKOS under the hood, TemaTres makes that possible.
What challenges have you faced in implementing TemaTres?
With TemaTres and SKOS, we now have the ability to create relationships between thesauri. This is a new frontier – external links have not previously been a part of thesaurus production workflows or thesaurus data. So, now we are busy linking legacy data, and revamping our processes and policies to create more interoperability. It is a lot of work, but the end result – the ability to extend major thesauri at the local or granular level – is tremendously powerful.
How do you see TemaTres and similar linked data vocabulary systems helping in the future?
The plethora of controlled vocabulary and ontology editors on the market allow us to publish not only metadata, but the organizational structures that underlie our metadata. This is powerful stuff for interoperability and knowledge-building. Why wait on the future? Get started now!
What do you think institutions can do locally to prepare for linked data?
There are two answers to this question. One is about preparing our data. Linked data relies on URIs and relationships. The more URIs and relationships we can squeeze into our data, the better it will perform as linked data. Jean Godby and Karen Smith-Yoshimura give some great advice on prepping MARC data for conversion to Linked Data. Relationships – that is, predicates in the RDF triple – can be sourced from relationship designators and field tags in MARC data. So, Jean and Karen advise us to add relationship designators and use granular field tagging.
The second answer is about preparing our staff. In the upcoming volume 34 of Advances in Library Administration and Organization (ALAO), I discuss training, recruitment, and workflow design to prepare staff for linked data. Library catalog theory (especially our tradition of authority control), metadata skillsets (to encode, transform, query, clean, publish, expose, and preserve data), and current organizational trends (towards distributed resource description and centralized metadata management) provide a solid basis for working with linked data.
Librarians tend to focus on nitty-gritty details – hey, it’s our job! But, as we prepare for linked data, and especially as we plan for training, let’s try not to lose the forest for the trees. Effective training keeps big picture concepts in sight, and relates each lesson to the overall vision. In the ALAO chapter, I discuss a strategy to teach conceptual change, inspire creativity, and enable problem-solving with linked data technologies. This is done by highlighting frustrations with MARC data and its applications, then presenting both the simplicity and rewards of the linked data concept.
Do you have any advice for those interested in linked data?
Do not simply publish linked data – consume it! Having a user’s perspective will make you a better data publisher. Try this exercise: Take a linked data set, and imagine some questions you might pose of the information. Then, try to construct SPARQL queries to answer your questions. What challenges do you face? And how would you change the dataset to ameliorate those challenges? Use these insights to publish more awesome data!
I want to thank Allison for participating in this wonderful interview. I encourage you to check out TemaTres and to think about how you can begin implementing Linked Data in your libraries. Stay tuned for the next interview!