Paper version of this classic presentation: http://www.cs.virginia.edu/~evans/cs655/readings/steele.pdf
Someone at a conference told me about this. I love it. It’s like a movie where you can tell something is not quite right, but you can’t tell what until the reveal.
This is my favorite programming talk that one can find online. I rewatch it frequently.
While I did like the first part of the talk, I was a bit disappointed with generalisation towards the end. While it would be great if matters of state and society could be simply expressed, and I am not saying they can’t be simplified or that doing this is bad, it’s often simply not possible, since these things aren’t designed or axiomatically derived. I tend to see this attitude quite frequently among engineering and programming circles, to dismiss the actual complexity of society, and for examine say “e-democracy will fix it” (one I personally find particularly distasteful).
The same point can be brought up with physics, where it’s even more obvious that a language “made” to communicate between everyday people about everyday activities and ideas, won’t be fit to describe the true nature of a universe so foreign to our everyday experience.
Thus, as people who work with computers, we shouldn’t want to reject at worst, constantly reduce everything to fit it into a program at best, but accept, understand and help where possible with these real complexities of the world
Less-code/faster/correct/durable are real metrics: they’re extremely well-defined and have very good business value.
But who here can’t see value in”readability” and “safety”? These are less well-defined, and at best, a function of how crap (average/most) programmers are at their job. They’re absolutely social, but they’re not a good way to value our work and ourselves since as more programmers become programmers, the mean (average) programmer simply gets worse.
Physics doesn’t have this problem, because the number of people who can tell the universe what to do, is much less than the number of people who can tell a computer what to do.
This could be one of the best presentations ever made.
I don’t think I’ve ever seen a talk by Guy Steele that wasn’t compelling and this is a true classic. He’s a great thinker and a really talented presenter to boot.
This is a slow burn but very delightful by the end of it.
I just wish somebody would do a parody talk reflecting, say, JS and the web. :P
What’s the take of those with a PLT background on this talk? cc @pylon @puffnfresh
Uhhh I watched it but I’m not sure what to think. I guess I don’t really get it.