peteg's blog

Trung Nguyên: 218B Pasteur, District 3.

/AYAD/HCMC/Cafes | Link

It struck me that I never wrote up this pearl of a Trung Nguyên café, even though I've been going there for ages. Loan pointed it out to me a long time ago. Darren rated the phê dá "not bad", as I recall.

The main attraction of this place is that it has the pentafecta [*]: good coffee, food, electricity, wifi, and is not too smokey indoors. I have to say that I prefer the cà phê sữa đá number 5 they serve me at 603 or 346 Nguyễn Trãi. The food is quite OK and not too expensive, and the courtyard is pleasantly shady. The parking bloke seems to be in a perpetual good mood.

[*] penta- (5) plus perfecta, cf trifecta.

Alan Chalmers: What is this thing called science?

/noise/books | Link

This book was on the reading stack for a long time; I believe I purchased it at Gould's many years ago. Unfortunately it happened to be the outdated second edition, without the additional, possibly fascinating, chapter on Bayesianism. I read this book as I've always been interested in the philosophy of science but never received any formal education on the topic.

I came away quite impressed by the first half of the book, where Chalmers takes an axe to naive inductivism and falsificationism. I was curious how these arguments relate to Ehud Shapiro's MIS, and machine learning in general, and came to realise that there the languages are quite rigid, with a careful identification of "observations" and "theoretical terms" that skirts some of the problems with refining theories in the face of unreliable evidence. It remains unclear to me how much one can learn about science-in-the-large from MIS, though the algorithms are cute beyond belief.

The latter half on research agendas, paradigms, programs, and the division of science into different activities lost me, largely as my interest in how a given scientific theory is structured and refined by "normal scientists" was unsated by the first half. The accounts of the higher-level activity of "disruptive science" offered by Kuhn and Lakatos are also interesting, of course, but stand on a different strata.

Samir reviewed the third edition. I concur with him that some discussion of what constitutes scientific explanation might have been helpful.