To make sense of the world, we partition it and put it into labeled boxes called "categories". Our brains do this automatically, and it is tremendously useful. It is what enables us to be effective agents in the world. It is also imperfect. Reality doesn't really fit in our boxes. This gives rise to the phenomenon of edge cases, which is a problem in software systems.
Being an edge case is an artifact of the boxes we choose, not of the things themselves. When we deliberately construct the boxes we use, we call it modelling. It's always a trade-off. What's a good box? How many boxes should we use? What happens if the number of boxes shrinks or grows? In this talk, we'll discuss theories of categorization, look at many hilarious real-world examples of what happens when our boxes fail us, and in general root for the platypus.
Being an edge case is an artifact of the boxes we choose, not of the things themselves. When we deliberately construct the boxes we use, we call it modelling. It's always a trade-off. What's a good box? How many boxes should we use? What happens if the number of boxes shrinks or grows? In this talk, we'll discuss theories of categorization, look at many hilarious real-world examples of what happens when our boxes fail us, and in general root for the platypus.
Einar Høst
Norkart
Einar W. Høst is an architect and developer at Norkart. He enjoys collaborative modelling, API design and computer programming. Over the past twelve years, he has done talks on a variety of topics, including hypermedia, resiliency, recursive art and lambda calculus. He has a PhD in Computer Science from the University of Oslo.