curiousgibbon

Who is "the industry"? My managers want me to succeed at the metrics. Research is too diverse for them to have real understanding of all of their underlings and their work so metrics become important. My peers (well, the ones I respect) know that the metrics are awful. The metrics seem to have the property that people who I consider successful do score well, but it's possible to score well entirely on the back of work with no scientific value too.

Except that for someone outside CS theory it will be hard to tell what is routine and what is ground-breaking. The proceedings for the recent editions of top theory conferences are here:

- FOCS (Foundations of Computer Science) https://www.computer.org/csdl/proceedings/focs/2022/1JtvJWOs...

- STOC (Symposium on the Theory of Computer Science) https://dl.acm.org/doi/proceedings/10.1145/3519935

- SODA (Symposium on Discrete Algorithms) https://epubs.siam.org/doi/book/10.1137/1.9781611977554

They're mostly paywalled but conference or full versions can usually be found on arxiv.org, and there are successful efforts to host paywalled papers illegally for the public good.

Personally, I am impressed by the progress on approximate counting and sampling. Developments include localization schemes (FOCS), average-case entropic independence (FOCS), the sampling/counting Lovász local lemma in (FOCS and SODA), and approximate counting with global constraints (STOC).

This logic can only make sense based on the axiom that almost nobody can be coerced into compliance. Sure, if fixed cameras are somehow impossible and the best hope for enforcement is "no more than 10 over no matter how bad the conditions" this may be the case. But in some countries other than the US these are solved problems.

There's a big cultural difference between the US and UK driving signage. In the UK there are fewer signs and it's generally upheld that the speed limit is a limit, and that what constitutes lawful driving may require driving much slower. You're doing to have a hard time suing anyone if you lose control of a car on a tight and narrow bend on a country road but were doing 59mph (i.e. within the limit).

Contrast the US where little yellow signs on every bend suggest a speed at which it's safe to go around, usually a conservative one. I think something missing from US road signage is the idea that more is worse: drivers are trained to ignore signs and speed limits because there are so many verbose and pointless signs around. It's much harder to trust the installers of the signs when they're obviously silly. In the UK it's more as if signs likely indicate some legitimate hazard that a driver will find out about very soon.

This may be an indication that it's at least useless, if not irresponsible, in other contexts too.

Aside from this craving violence angle, which I don't take seriously at all, are there any redeeming qualities of American football compared to other sports? I guess the unlimited substitution thing makes it participatory at e.g. high-schools level as teams can be enormous. But a lot of expensive equipment is required and the brain injury problem seem to more than compensate for any possible benefit of getting kids to run around. Soccer is clearly a rotten affair internationally (FIFA), but the sport itself seems much less harmful on several levels.

Europe has had instant, free bank transfers for years without any crypto in the underlying tech. The same regulators taking away crypto are perpetuating the horror show that is US retail banking in other ways.

This is an active area of research for sure. One idea is basically to come up with a few rules for what the districts must look like (as unions of precincts) and use Markov chain Monte Carlo to sample from a probability distribution on the possible districting plans. Several court cases have featured expert evidence from math researchers asked to quantify how crazy the districting is using this methodology (ish). Wesley Pegden's expert witness report in one such case is readily findable, for example.

One of the main difficulties is that we don't have much rigorous theory telling us how long to run the Markov chains.

In other countries there are fixed cameras, which can even be programmed to have some grey area (do they trip at x+2, 1.1x, ...?). These can probably pay for themselves and then some if placed correctly. They're conspicuously absent from the US.

I agree, of course metric is better. My point is that perhaps things can improve in a direction that is orthogonal to the metric vs imperial/US choice and doesn't require making anyone angry.

The inherent superiority of metric is small enough that people are going to be angry about this stuff, and yet it's dwarfed by inherent idiocy of inconsistent labelling which is solvable without a prolonged culture siege.

I've dealt with the British and American version of imperial measurements, and know metric. It's not exactly difficult to quantify and convert as necessary so all the culture wars confuse me.

What enrages me about the American system is more the lack of consistency. Take something like mayonnaise. One brand will be sold with a price quoted in dollars per ounce, another in pounds, and yet another in x$/yoz for some y other than 1. And one will have calories per 2oz portion, one calories per 7oz portion, and the other calories per tablespoon (0.76oz) on the label. Obviously I made the numbers up, but this phenomenon is pure insanity. And it's blatant that the obfuscation is to enable exploitation of the consumer who has to perform arduous arithmetic to compare products. At least in the (pre-Brexit) UK the yoghurt had to have calories and price per 100g on the label, whether it was a 454g (vestiges of 1lb) pot, 500g, or some other size. I prefer consistency in some unit that I can learn to chaos.

As for recipes: why insist on measuring things of totally unknown density in cups? Mass is the appropriate way to measure strawberries or lettuce leaves, not volume.