Supervising a PhD student is a learning task for the respective supervisor. Granted. But it’s also a reminding task as I was reminded of (no pun intended) this morning.
Björn Forcher, a colleague and PhD student I supervise on the topic of explanation, discussed with me his current ideas on his research work. He talked about what could be explained and pointed out that not everything can be (easily) explained by humans. So we should not expect an information system to be able to explain everything and easily. This observation is very much in line with my human-centered approach to (explanation) research (an observation that tends to slip from my mind regularly. Well …). This again reminded me of an interesting discussion I had with the students of my last Semantic Web class about identifying things and URIs resulting in the observation that people do have problems in everyday life with identifying things or people. How often do you talk with someone about someone else until you realise that both of you talked about different people? So why should information systems should have fewer identification problems?
My point here is: Think about computational problems always also in “real world” terms. One might get surprised.
[composed and posted with ecto]