shub

Conspiracy theory? How so? We know that Russia spent some amount >= $100,000 on campaign ads, and guessing the actual amount comes down to how good you think Facebook is at resolving accounts to people who don't want to be found. I don't think it's completely unreasonably to think that the actual amount of Russian spending was on the order of $1-10 million.

Iit's specifically minor bribes required to get anything done even if you comply with the law. Bribing a building inspector to overlook safety problems, that's an FCPA violation. Tipping him $20 to thank him for his hard work on verifying your construction's code compliance, not a problem. The text is so:

> Subsections (a) and (g) of this section shall not apply to any facilitating or expediting payment to a foreign official, political party, or party official the purpose of which is to expedite or to secure the performance of a routine governmental action by a foreign official, political party, or party official.

1933. NSDAP was in the Reichstag for a while before that but not really involved in governing.

It doesn't. AIUI jlink creates a binary image of the VM + referenced classes. AOT is a lot easier to do with modules implemented and I'm sure it's on the roadmap, but not available yet.

So the input log isn't partitioned? I'm going off of https://fauna.com/blog/consistent-transactions-in-a-globally... which states

> Each database has a single partitioned RAFT-replicated log that is used to derive a total order of all transaction effects, and that handles all cross-region network communication.

But if each partition knows about all the reads in a transaction and verifies all the timestamps, that sounds workable. Not that I'm in any way qualified to pass judgment on something like this, but I feel like I have a little bit of understanding. Overall it sounds really cool and the idea of agreeing on inputs instead of outputs is tremendously clever.

Except that Germany 1933-1939 didn't have significant inflation.

How does FaunaDB guarantee that all partitions of a transaction commit or abort? The Calvin paper just says "see here for our distributed pessimistic locks that make this not a problem" but if FaunaDB uses oplocks then that doesn't apply.

But there is AGP CD-ROM CC0 QQZ, so, you know. There's that.

Well, yeah. Stalin was a monster. He was just good at being a monster.

I wouldn't say that, actually. Germany's recovery was primarily fueled by massive deficit spending on rearmament and social programs. Keep the people pacified, bootstrap from the tiny military allowed by Versailles up to something world-class, then take over Europe and Asia. Not sustainable at all, and I think it's really debatable whether Hitler deserves any of the credit anyway. His role was more or less to turn on the money tap and tell the Prussians to tool up, which they did very happily and effectively because they were Prussians.

That quote sounds good, but I don't think it's necessarily applicable to this situation. The author seems to be talking about complex systems that are designed and operated to be robust against failure, like the space shuttle. Saying that Challenger blew up because of an O-ring is technically correct but also horribly wrong, as an example. Equifax IT does not appear to be operating at a level to prevent a single failure from causing terrible damage all on its own.

That aside, it's hardly true that one person can bear all the blame for not patching their systems, even if they did successfully prevent patches from happening. For one thing, how the hell did they keep their job after doing that? Unless it was the CEO (well, now that they have a new CEO maybe they'd like to put all the blame on him), there was someone up the chain who could insist that the patch get applied. I think you definitely could apply root cause analysis techniques here, and I strongly suspect that such analysis would uncover numerous serious deficiencies in Equifax's IT operations. Of course, guessing that a large boring corporation has terrible IT practices is similar to guessing that a given duck quacks and has wings, so there's that.

Where people seem to really care about perf is networking. What that means in non-crazy land is you cut out the kernel from that part, do networking in userspace, and pin a thread to a CPU to poll for packets. Like it's all MMIO and DMA anyway so ring 0 doesn't have to get involved if it doesn't want to. Then for other stuff you have convenient slow OS facilities. Including threads, processes, and disk I/O! Luxurious.

So it's not just that this has to perform better than a kernel, it has to perform better than not involving the kernel in the critical path, and it has to perform so much better that it's worth wanting to beat yourself in the face with a hammer after trying to diagnose the latest lockup. Unikernels are interesting to me, but like the demoscene or a semi tractor doing wheelies. Not for production usage.

I have 5 tables which have overrun the range of int in the id column. Happily they were defined bigint. A while ago one ran out of tinyints, but doing the schema change...was pretty terrible actually due to the foreign keys. The tinyint decision wasn't me, I want to make clear. New tables all get bigint. Either you don't have many rows, so you can spare the space, or you have a lot of rows and the thought of changing the type of your primary key gives you a queasy feeling in your belly.

That's true. Thanks for reminding me! Next time I need to torture a graph store I'll try loading common crawl into it.

I was referring to x1e.32xlarge..2x1920GB ephemeral, 25 Gbps Ethernet + 14 Gbps EBS Ethernet. And at least for the application I'm considering, the working set would fit in RAM and you can run the algorithm online. Score!

That's what I mean. You have to get really big to have Big Data. Fine, you can't do PageRank on a petabyte of web crawls using one machine. But the datasets that people use for benchmarking, at least the benchmarks that are made public, you definitely can. You can go far larger than a typical benchmark dataset and still do it on one machine.

I had cause to consult a syscall table (not this one, a correct one, forget where I found it) when doing something or other with fasm. fasm's macros are pretty dang advanced...I remember having argument length checking and rudimentary type checking as well. Then I got done yak shaving and remembered that programming in asm sucks.

But yeah, it's for writing your own syscall wrappers. Something not exported by libc, or more likely if you're not using libc.

Good luck finding anything public about graph processing on a dataset too large to fit on a single machine. I can launch an AWS instance with 128 cores and 4 TB RAM--how many triples is too many for that monster? Tens of billions? Hundreds of billions?

Yeah, that's me in a nutshell. My employer recently sent everyone new hardware and I was initially bummed about getting the just-before-Touch-Bar MBP. Then I saw how gimmicky and awful Touch Bar was for the "lucky" colleague stuck with it. I'm typing this on a five year old MBA that isn't quite up to scratch for work anymore. If Apple would take this case, stick updated hardware and a Retina display in it, I'd buy it on launch day. Never gonna happen so I guess I'm never gonna buy another Apple laptop.

Desktop Linux, especially when your employer runs Exchange, is a barrel of pain I've opened far too many times for my own good. Not going that route again.

IANA neuroscientist but it seems reasonable that the brain will fix up small errors. It has to anyway, because it has its own input latency. Besides, the article is referring to discrepancies between multiple audio tracks, which don't depend on vision at all.

For a real-world example, listen to 2 TVs several meters apart and tuned to the same channel. At least with OTA or cable you can expect them to be playing ~simultaneously but the skew between the received signals is easily perceptible.

Don't they know that money is bad for them?

On some systems delete is backspace when you're at the end of a line.

But we already have that. It's called practicing.

Sounds like someone wasn't practicing defensive electrical engineering (is that a thing? I think it should be a thing).

Everything old is new again! Including, like, horses pooping on muddy streets.

Should a neo-Nazi have phone service? Would it be acceptable for Verizon to cut off someone's FiOS based on the Anti-Defamation League's say-so? For Internet-based businesses, the ability to display ads and process credit card payments is nothing more or less than the ability to exist. It scares me that these decisions will be made by trade groups and multinational corporations, largely or entirely free of judicial review.

We're talking about people having their work destroyed and possibly seeing their livelihoods ruined because they harmed a company's profits. I'm not comfortable with that, and I'm sure as hell not comfortable with it being done as easily as the article suggests.

Let f(x) = y. The guards have f and y. A spy gives a guard an x, and the guard computes f(x) and checks it against his list of ys. If it's on the list, the spy can pass. The enemy has f and y too, but it doesn't help them! f(x) is trivial to compute but the inverse is hard, so the enemy knows exactly the answer they want the guard to get and no idea how to make him get it!

One-way functions are very cool and form the basis of public-key cryptography, although it's quite a bit more complicated than this example.

I've seen doors kicked in where it was the jamb that broke and not the lock. I think I'd rather call a locksmith...

The way I see it is, if UK2 wants to fuck customers out of the service that UK2 advertised and the customers paid for, the customers are welcome to fuck right back.

To make a long story short, IE9 is fast but very beta. Now you don't need to try it!

All that ever happens is that you "call the functions and pass-in the data", except that OOP languages know to pass this for you. The difference between OOP-style C and an actual OOP language in this regard is syntactic, not conceptual.