Today I was reading through (again) one of my favourite PhD Theses of all time (not mine!). It was titled: A Dynamic Learning Object Life Cycle and its Implications for Automatic Metadata Generation by Kris Cardinaels from KULeuven. This one, along with Xavier Ochoa’s, Learnometrics: Metrics for Learning Objects and of course mine, are my three top PhD Theses… It makes sense as they are all about metadata..!
Reading through the Thesis, I started thinking about metadata, about learning objects and their lifecycle through a repository’s lifecycle. Then I started thinking about libraries, preservation and librarianship in general, which kind of opened a bir and wide circle of things in my mind, so I decided to document them here as food for thought and future reference.
I am NOT a senior in the field. I am not a guru either. I consider myself somewhat of an expert on metadata that still has so much to learn and grasp. Having said that, I can now mumble as much as I want without claiming any higher wisdom or epiphany! I have read bits and pieces of the ongoing research from 1995 onwards. This is research that was carried out while I was in my room, playing with my first 486 desktop PC, not even Pentium (!), attending junior high in Athens, Greece, at the age of 12, and is still carried out.
Cardinaels offers a really interesting approach that kind of reminded me of the theory that librarians go about for years now. He introduces the dynamic life cycle of a learning object which is a fundamental issue we have to agree on before we move forward. I think it pretty much resembles to the need of keeping learning resources up to date, preserving them and their descriptions (call me metadata). Librarians have long struggled with such things and it’s kind of amazing how little convergence you see in related research between information scientists and librarians.
I am all for a learning object that evolves through the time, aquiring automatically generated metadata, starting from the simplest (size, duration, etc.) moving to the most complicated ones, like the Educational elements or even classification terms. The premise of picking metadata out of the world wide web for each learning object, based on context and use from various actors, is mind blowing indeed and will solve many many problems. Take into account the FRBR-ization of the Educational Metadata Standards coming our way and you have something there…
I have been a supporter of curated, high quality metadata that come mainly through human annotators. I still am in a way. I based all my PhD research on that you see, so I kind of have to stuck with this, at least for a little while after finishing my PhD! 😉 Apart from that, I still think that human annotators are crucial to the entire process and I am a firm believer of the fact that we need to support them as much as possible through automated means.
On the other hand I am thinking about our education, about our online schools and repositories. And on a philosophical level (I am Greek you see), I think that we need to have some kind of human-generated content involved. Some kind of effort. Automated metadata and automated content are nice, but they lack the “intent”, the “master design”, the “rationale”. We can automate-away as much as we want. Heck, we can outsource all the process to an automated algorithm that creates learning objects out of learning assets (I heard about this recently and I also like it a lot!) and adds metadata while she (don’t know why, it just seemed like a “she”) serves the finished product to the learner.
Being more of a traditional kind of guy I guess, I prefer a small repository with curated content for a specific community of users, trusted and well-defined. With automated metadata when possible, with powerful content design software and all the goodies someone can ask. I am impressed by aggregations of content but I don’t fancy them too much. Millions of resources are OK but if you look at some of the really famous mega-repositories, you won’t find too much quality, neither in content nor in metadata. Cheap is one way to go in educationa, but sometimes, too cheap means also useless in my opinion.
What strikes me most of all, is the fact that although we seem to want to move away from Google-like searches, in a way we end up aggre-googling our nice repositories into bigger ones. And on the other hand, it seems like a waste to have all these possibilities in our hands and not go for something bigger…
So, still searching for the balance between everything!