Comments for blog :: Brent -> [String]

Comment on Workshop on Functional Art, Music, Modeling and Design by Brent

Brent — Thu, 16 May 2013 10:51:17 +0000

Thanks, fixed!

Comment on Workshop on Functional Art, Music, Modeling and Design by Harold

Harold — Thu, 16 May 2013 04:20:37 +0000

On the website

http://haskell.cs.yale.edu/people/paul-hudaks-home-page/

should be

http://haskell.cs.yale.edu/people/paul-hudak/

Comment on Beeminding for fun and profit by Gábor

Gábor — Wed, 08 May 2013 16:38:57 +0000

Right, if the primary goal were to be productive, the logic would be unimpeachable. But the primary goal is to not be miserable, while “not sucking” is a secondary goal. If Beeminder makes me more productive at the cost of being more miserable, then that’s not a good trade.

But as I replied to Brent – theorizing can only get you so far. I’ll have to try it out and see for myself. Once I stop procrastinating on it.

Comment on Beeminding for fun and profit by Gábor

Gábor — Wed, 08 May 2013 16:32:19 +0000

The more I think about it the more not-so-simple it all looks. The nagging sense of guilt is familiar, but if I just ‘drift’ and don’t pay attention to the things I “should be doing”, I feel basically fine (unlike your case, where apparently that exacerbates it). If I really try to push myself, in some cases that can push the stress and anxiety level way up. While in other cases I just need a push to get started, and after that I’m actually enjoying myself and not just drifting.

So I guess it’s futile to try to predict how well Beeminder would work for me, and I’ll have to get around to determining the answer by experiment.

Thanks for sharing your thoughts!

Comment on Random binary trees with a size-limited critical Boltzmann sampler by Brent

Brent — Wed, 08 May 2013 11:35:02 +0000

Hi Jonas, good idea! That had not occurred to me. I will give it a try. You’re right that splitting would be essential to make this work — and while people have doubts about the theoretical properties of splitting that probably only matters for cryptographic applications.

Comment on Random binary trees with a size-limited critical Boltzmann sampler by Jonas Duregård

Jonas Duregård — Wed, 08 May 2013 09:47:06 +0000

Nice post Brent!

One comment: wouldn’t it be nicer to keep the simple tree generator and just use a lazy size function and lazy size comparison to sieve out the large (and small) trees? For instance size could return a peano number or you could just make a specialized function smallerThan :: Tree -> Int -> Bool.

One potential pitfall is that the simple generator is not as lazy as it could be (can be fixed using a splitting random number generator).

Comment on More counting lambda terms by Pierre Lescanne

Pierre Lescanne — Fri, 03 May 2013 12:42:52 +0000

I published with Katarzyna Grygiel a paper entitled “Counting and generating lambda terms” http://arxiv.org/abs/1210.2610 which addresses in particular the problem of counting simply typeable terms. In particular, notice the sequence https://oeis.org/A220471 on the Online Encyclopedia of Integer Sequences.

Comment on Enumerating linear inhabitants by Monad transformers: a cautionary tale | blog :: Brent -> [String]

Monad transformers: a cautionary tale | blog :: Brent -> [String] — Mon, 29 Apr 2013 21:07:04 +0000

[...] aren’t (in general) commutative! Think carefully about what order you want. (I actually wrote about this once before; you’d think I would have learned my [...]

Comment on Random binary trees with a size-limited critical Boltzmann sampler by Monad transformers: a cautionary tale | blog :: Brent -> [String]

Monad transformers: a cautionary tale | blog :: Brent -> [String] — Mon, 29 Apr 2013 21:07:00 +0000

[...] ← Random binary trees with a size-limited critical Boltzmann sampler [...]

Comment on Random binary trees with a size-limited critical Boltzmann sampler by Chris Warburton

Chris Warburton — Fri, 26 Apr 2013 12:21:45 +0000

Oops!

size x = 1 + map size (children x)

Should be:

size x = 1 + sum (map size (children x))