You can also reproduce it within a week without hosted cloud services. What matters is that you don't have to develop custom software and instead spend that week writing config files and orchestration scripts, be it cloud stuff, docker containers or whatever.

I can reproduce it without cloud services sure. But then I have to maintain it. Make it fault tolerant. Make sure it stays patched and backed up, buy enough hardware to make sure I can maintain peak capacity instead of having the elasticity, etc.

I have done all of this myself with on prem servers that I could walk to. I know exactly what’s involved and it would be silly to do that these days

So they assume four-dimensional space-time of a certain shape, similar to how a two-dimensional sheet could be curved taking various shape. Then they calculate how would physical objects behave in space-time of that shape.

Can such shapes exist anywhere in our universe, realistically or even just theoretically? For all we know, perhaps not.

I can name several that can cover significant part of local demand with solar power. Or perhaps they'd have energy storage to fully cover it by now had they invested into that.

Depends if the project started with you or was ongoing. If it is the latter, defer to pre-existing coworkers to an extent at first.

For chronic exposure, yes. You don't want it to be in your food all the time. But accidentally eating a tiny bit once would be safe.

That’s the whole point of this article. Many people will consume multiple servings of protein shake per day for most of their life, and they’re contaminated.

They're contaminated at levels comparable to other foods people eat (baseline).

No they're not. If you're using an array with length over a billion in Java, your code stinks already before you start using binary search.

That's not only wrong in itself, but totally orthogonal.

A binary search implementation should still work, regardless of the array length, or have the limitation documented.

And of course an array "with length over a billion" can be totally valid, depending on the use case, your tradeoffs, available memory, etc. It could even be the optimal data structure for some use cases.

I'm not a Java programmer, but how would you load of a 1GB file into memory? I assume read returns some kind of an array.

Also big arrays being (supposedly) a coffeee smell doesn't mean that code handling them improperly is not buggy.

If you really needed it in memory you’d use one of the file APIs that will map it and present a direct byte buffer view over that memory.

Those APIs use long as their offset unlike the 32 ints used by arrays, and would avoid having to copy the data into some other object.

There has been some discussion over the years about how arrays could be changed in the JVM to support longer lengths, but doing so without breaking existing code and while providing truly useful functionality without providing obvious footguns isn’t as easy as you might think.

Java's arrays use a signed 32-bit int as their length, so the longest they can be is about 2 billion elements.

If your code has arrays over a billion elements, then it will fall over the moment someone inputs slightly larger data

Relational databases often require searching and sorting gigabytes of data to answer queries (sometimes larger than RAM if e.g. k-way merge sort is used) so it doesn't seem that far-fetched, especially given that there are database systems written in Java.

not doing much scientific programming eh?

I've driven electric motorbike for 7 years and it's absolutely great for everyday commute. Lighter ones tend to have removable batteries so they work even when you don't have a garage with electric outlet. Some heavier ones you can even take on an occasional long trip, though that's not super convenient – it's commuting (combined with joy of riding) that makes them useful.

One does need to know where one is going to service it, though, because they can sometimes have stupid electrical issues which are objectively easy to fix but hard for you to fix on your own cause you don't know which wire goes where.

It does seem to play worse than it should, by game went: 1. d4 d5 2. c4 dxc4 3. e4 Qxd4 4. Qxd4 5. Bc4 6. Nf3 7. O-O 8. Rd1 9. Qd8#

Oh no dude, it's worse than that: c4, Qa4, Qxa5, Qc7, Qxc8# lol

Nice find. I tried to get an even shorter mate, but its tendency to give up its pieces got in the way: 1. e4 e5 2. Bc4 Bc5 3. Qh5 Bxf2+ 4. Ke2 Bxg1 5. Qxf7#

Can anyone get a mate in 4?

Oh, that was easier than expected: 1. e4 e5 2. Qh5 a6 3. Bc4 a5 4. Qxf7# I assume a mate in 3 (or a Fool's mate in 2) is not possible...

1. e4 e5 2. Nf3 Nf6 3. d3 Ne4 4. dxe4 a5 5. Ne5 b5 6. Qf3 a4 7. Qf7#

Still, impressive for "a pile of regexen".

If one divides weight of the airplane by the number of wheels it has, one would find one wheel carries around 10 times more weight than that of a truck. You even get a slightly better deal on landing when you don't have as much fuel.

That's a lot of weight but nothing crazy, so on a dry runway wheel brakes alone are more than enough to stop normally. They would also wear out a lot, overheat and occasionally ignite if used like that, so that's what thrust reversers are for.

That mass has much more velocity than a truck.