busy day, formal philosophy workshop, boat trip; new housemate; Nelson in India

Busy day meeting lots of new people. Met many new students, some of whom seem interesting.

Branden Fitelson, who is visiting Amsterdam, is also a very interesting person. I name-dropped easwaran, and he said that his talk had passed the Kenny test.

---

Edgar moved in for 2 or 3 weeks, until Nelson gets back from his gig in India, fusing Brazilian music with Indian music: cavaquinho + guitar + sitar + tabla (btw, I was very impressed by the way he got that, just making a few phone calls)

Fitelson diagnoses my philosophical view

me- aren't all these philosophical questions, questions about AI? (unifying epistemic logic with probability, coherentism, etc)

somebody- what about metaphysics?

me- irrelevant things are, well, irrelevant.

Fitelson- early Hempel meets Herbert Simon!

me- ...uhmmm.. ok...

I haven't read any Hempel.

housemate is colorblind

I just found out that my new housemate, Edgar, is colorblind.

I've had tonedeaf housemates before, but never a colorblind one.

I want to find the right shades of red/green that he can't distinguish... he says that he can always detect boundaries, but I suspect he's wrong. On his side is the evidence that pictures in colorblindness tests are boundary free: they make the number in a swarm of dots...



My theory is that colorblindness can be modelled by a projection from a 3-dimensional into a 2-dimensional color space R x G x B -> RG x B, where the red and green dimensions collapse into one red-green dimension. (oh, pardon me, I'm being 3Dcolor-centric: the real color spectrum is (potentially) infinite-dimensional)

In practice, I think colorblind people often distinguish green from red by darkness levels... but this obviously fails a lot.

I think they make the test this way (boundary-free, dots being different shades of red and green) because different individuals have different projections. So for the test to catch most red/green colorblind people, it needs to be this way.... although they could also make a test with tiles instead of dots.

Also, note that the thickness of the numbers is 1 or 2 circles. If there were many more, then some colorblind people (those whose projection is far from the average projection) might notice a statistically-significant difference in darkness, and therefore see the numbers.

Oh, Henry Sturman has convinced me that many colorblind people are actually differently-color-visioned.

(Pardon my inconsistent UK-US spelling.)