After two years of Corona we had the chance to visit ELAG in person. ELAG 2022 was held at the National Library of Latvia in Riga, which is a very impressive Library. ELAG is known for being a forum for exchanging knowledge about new library and archive technologies, as well as connecting people working in libraries.
The keynote was held by Andrejs Vasiļjevs about “Artificial Intelligence – to serve or to be served?”. He explained how AI can help us in our everyday lives, but also how libraries have an important role to play when training AIs. It is important to have high quality data when teaching an AI, and Vasiļjevs highlighted how one of the new roles of libraries is to serve this need.
After the keynote we had the chance to present our findings. We compared different triplestores according to various attributes. We explored subjective aspects like, “how easy is it to install and run the database?”, “how useful is the documentation?” and “what restrictions does the license have?”. We also compared some performance aspects like, “how long does it take to perform different queries?”, “how much ram is necessary?” and “how much disk space is used?”. The most interesting finding, is the different query times when using SELECT or CONSTRUCT queries. SELECT Queries can be over twenty times faster. We suspect the different serialization methods used when executing various queries could be the reason for this result, but we will analyze this further. For further details see the presentation slides below.
Next up, Christopher Bartz from ZBW talked about “Software Architecture for the Automatization of Subject Indexing”. At ZBW they automate their subject indexing due to the huge amount of digital documents released every year. They use a microservice software architecture which periodically fetches metadata to index it. These descriptors that are created in this way are then stored in a database that feeds their search portal.
Then, Jakob Voss presented “Data Validation made easy”. One of his main points was that non-trivial metadata is inherently dirty. He introduced an open source validation service to validate data against different languages and formats.
Finally, Anna Russo spoke about “Bridging the Gap: Enhancing Digital Discoverability of Library Resources & Services”. She pointed out that most students start their search for information on websites like Google Scholar and don’t use the tools from their own university libraries. She then went on to explain how libraries can increase the discoverability of their resources and services.
We thank the National Library of Latvia for the wonderful conference and the excellent catering very much. We hope ELAG will be in person again next year.
Elag-DBs