The slightly concerning thing is that the notice email I got was in my Spam folder. I checked the source carefully and the password reset link appeared to be legitimate. So I've used it (entering my email address only). The next email was also marked as Spam, with GMail saying that a lot of mail received from postmaster.scribd.com is spam.

Has anyone got any thoughts on this? Has scribd done something dumb in the past? Has their mail systems been compromised too? Is there a concerted effort to fool GMail into treating these password emails as spam?

Isn't the N-body problem already a stronger result than this?

Or are they orthogonal? I'm interested in any explanations about how they compare.

Disclaimer: IANAP and have not RTFA.

Your "basic logic" has missed the part where being unable to formulate a trend for a 10 year period != having no access to the hundreds (ranging to millions, for some measures) of years of data that we have.

The GPs point was that while we may may need more than 10 years of data, we do have more than 10 years and we can draw trends from them.

Measure what is measurable, and make measurable what is not so.

-- Galileo

We should combine that with one one of Einstein's to make the pragmatic rule: Everything should be as measureable as possible, but no more. (Basically, we should try to make things measureable, but we shouldn't measure things that obviously aren't.)

That is not the alternative -- it's the opposite extreme.

Allowing a media free-for-all increases the risk of Jurors getting outside information on the case.

Justice needs oversight; but also also needs to be protected from interference. Optimising justice means finding a tradeoff between these ideals.

Ok, correct me if I'm wrong here, but isn't Tsar Bomba famous for being the largest nuclear weapon detonated? I don't see the relevance of it to estimating the consequences of an accident at a power station. The effects in each case are almost entirely incomparable. Sure, they're both "nuclear", and each involves a release of radioactivity. But the distribution of that in terms of isotope mix, time, intensity, location follow entirely different models. Furthermore, Tsar is renowned for its fusion detonation, which AFAICT is largely unrelated to the amount of fissile material required to trigger it -- for all I know the fission bomb component was no larger than average. Using it as the benchmark for "biggest nuclear thing ever" is bizarre and simplistic.

Interesting. I'm not going to "fisk" that page (since fisking is a retarded practice that amounts to cherry-picking easily criticised minor points).

I was kind of proud to see my own local paper the "Wellington Dominion Post" scored a 7 for "selecting a picture of a mushroom cloud like explosion because they couldn't think of nuclear in any other terms than a mushroom cloud". Well that's kind of subjective: it doesn't look especially mushroomy to me. But it does look a hell of a lot like an actual Fukushima explosion photo.

There is a lot of sensationalism coming out in the Fukushima reporting. But sites like this aren't interested in accuracy; they exist to say that any concern for the plant is overblown, and to discredit any negative reporting of it, regardless of veracity.

ISTR (from an abortive combinatorics course) that there's a commonly used encoding that provides 3 bits of ECC for each 8-bit byte. Could this string be 3 simple bytes with ECC?

I'm a little unclear about where exactly this works in PostgreSQL. It's true that, for stored procedures (functions), there is a plugin system for the stored procedure language. But there is only one SQL dialect that can be used outside of stored procedures. The new DO command arguably expands this, but it still looks a lot like a stored procedure body.

Tiny quibble: I don't know if you could call it the "primary language" since it's just as primary as all the others. I recall a time when I had to enable it in each DB that needed it -- perhaps if it's included by default it is now slightly more primary than some others...