It seems that one of the issues that has been around for several years in the library world, the one of cataloging items, is here to stay.
I remember ten years ago the provocative statement by Erik Duval: electronic forms must die! This is a slogan that we all liked, adopted and worked for. So, a decade later, did we manage to exterminate them? Not yet, I am afraid…
Every time I hear about a new initiative that wants to help people organise their digital material in some kind of database, I know the type of question that will come next: it’s always a variation of “how can we explain to these people how they will describe (tag; annotate; index; catalogue; choose your verb) their materials so that they can manage and discover them in an easier way?”. And I know that after such an initiative launches, at least two questions will come up as well:
– How can we help people understand how they should describe their material?
– How can we improve the quality of the descriptions so that the serve the purpose of better management and discovery?
These are the practical questions that Nikos Palavitsinis transformed into research ones and investigated during his PhD, that I had the pleasure and honour to co-supervise with my good friend Salvador Sanchez from the University of Alcala, Spain. Having just watched Nikos successfully defending his PhD, I think that I can better reflect on these questions.
Generally speaking, Nikos has proven that we can indeed help people improve the quality of the descriptions that they provide. But this is not an easy, quick, or automated process. It requires lots of manual, hands-on, caring and meticulous work by the people managing the repository. Our domain experts are not born with information science skills. Our librarians are not born with domain understanding and knowledge. Our tech teams cannot automatically extract the human knowledge (yet?) so that they automatically do this job. And as poor metadata quality has a cost, so does good metadata quality: thanks to works like the one that Nikos did, we have a more structured and proven process that can be applied in different repository set up and population initiatives to improve metadata quality in measurable manner; and we can budget the cost of the application of such a process so that we can be aware of how much such a process will cost – compared to the expected gains.
I will not go into the details of Nikos’ process, but his presentation reminded me the power of deploying solutions that are adapted to the needs of real people and that are applied in a way that will not disrupt a lot their regular workflows. Nikos did not try to measure the quality of the metadata that people provide, and then get back to them to complain or instruct about improving quality. He rather developed a participatory and open process that involves them very early, enhances their understanding, shows them what good and bad metadata looks like, engages them in peer reviewing and appreciation activities. And at the end, he has shown that in all three repository networks where he tested the process, metadata quality has improved indeed.
I am really looking forward in putting this process in place for more repositories that we work with. If we cannot completely kill electronic forms (yet), lets see how we can make them lest annoying and disruptive for our users. And there are more ideas about how this work can be taken further:
– Let’s find ways in which metadata creation is a playful and creative process, rather than a boring form completion task (see more in this paper).
– Let’s take advantage of existing tools and services that can help us measure and discovery quality problems in very large digital collections, instead of relying on manual inspection or (even worse) end user complains (see more in this paper).