yosh

Yeah, this is Erlang:

    fib(0) -> 0;
    fib(1) -> 1;
    fib(N) -> fib(N-1) + fib(N-2).

Having recently run into some non-obvious crappy performance characteristics with I/O in Erlang, I'm curious if you have any pointers to docs or code that shows how to use Erlang I/O correctly so it's fast?

I'm curious, why would you write that game in Node now, if you already know Erlang?

From my perspective, Erlang is a better fit than Node for all the problem domains that Node is good at. I understand using Node if you don't know Erlang, but do know JS, but if you're already skilled with Erlang...

Probably Musion Eyeliner:

http://en.wikipedia.org/wiki/Musion_Eyeliner

Weekends are forced vacation too. It'd be great if companies adopted a "floating weekends" policy too, for the same reasons I outlined, but no company I've come across seems to do so, officially (you do see unofficially here and there bosses who recognize an employee working the weekend and allow them to take a couple days off later, and simply not reporting the vacation).

As to your first question, yes I've experienced both. Not month long shutdowns, but week long ones, but it's the same thing: if it's not convenient for you to take a vacation during that time, it's pretty much a waste and it sucks. If your spouse doesn't get 4 weeks off on August, are you going to go on a 4 week vacation then? Probably not. But then you'll resent the people who do actually take that 4 weeks off to go somewhere. It's bad for morale.

I agree that a solid 4 week vacation is enriching, but just declaring an arbitrary 4 week shutdown is a cop out. Fix the structural problems that require every one to be in the office at once, and then give everyone that much vacation time to be used when they like it. I'm not saying that this is easy to fix, but if you really believe in giving people time to have significantly different experiences in their lives, this is what you'd have to work at.

Forced vacation isn't freedom. If August isn't a convenient time to take a vacation, and the employee would rather work, let them work and take that chunk of time off at some other point. If taking that time off another time during the year is unacceptable, then your policy simply isn't fair to all your employees.

Where I work even the national holidays are floating holidays, so if I want to work on Christmas that isn't a day off for me, and I can trade that for a day off for any other day of the year.

Especially for travel, it's way cheaper to do so for off-peak seasons, so forcing August or the last week of December etc. when just makes for more expensive vacations. This is one reason why I think the "French model" kind of sucks.

There's actually 5 charging stations that Tesla installed on the 101 between SF and LA. If we got serious about electric cars, higher charging station density would help with long trips.

And I agree, SF<->LA is not a long drive.

FWIW, the longest stretch of interstate without services is I-70 in Utah, at 106 miles. I don't think I-80 has any >100 mile stretches without gas, but possibly 80 mile stretches.

Off interstates though, the west can get pretty empty. This is near the Nevada/Oregon border:

http://i675.photobucket.com/albums/vv114/cherokeeprogressive...

Reasonably good presentation, but this stuck out at me:

> One team that started with Erlang and moved to Java so they could find programmers. Think about those scenarios.

If the people you are hiring can't pick up Erlang in a couple of weeks of being on the job when they are doing that full time, with colleagues around who already know Erlang to help them, you are hiring the wrong sorts of people. Same goes for any other language/technology stack.

I've also used private ETS tables for this.

Kayak's got them now actually, but only schedules, not prices.

git ls-remote will help you to get the hash of HEAD. It's much cheaper than doing a zero-depth clone.

Yeah, here's a paper on that topic:

http://www.sussex.ac.uk/linguistics/documents/where_do_mama2...

I've noticed that sorted sets have noticeably more overhead than the other types.

T-Mobile is fine in the mission in my experience.

http://www.thinkwiki.org/ is a good resource for figuring out Linux compatibility.

I also recommend Thinkpads. The T and W series are good too.

Careful with casting strings to ints like that, it'll break on architectures that require aligned accesses, i.e. if you try to load a uint32, the address of the memory location must be a multiple of 4. x86 happens to be fine with non-aligned access, but many other architectures aren't.

If you look at the nginx source, that #define is guarded under NGX_HAVE_NONALIGNED and there is a fallback that does straight byte by byte comparison.

> No security check, no long waits, no cattle herding.

Not all trains. With Eurostar (Paris<->London) there's a security checkpoint with an x-ray and metal detector, along with passport control.

Yeah, but tweets aren't limited to the GSM 7-bit character set, so if you tweet in say, Russian, 140 characters is too much, because the whole message needs to be encoded using 2-byte characters.

Twitter's limitation is based on thinking around SMS, but it doesn't actually match up to real world SMS, because they didn't think through all the international language considerations. It's pretty botched up.

See http://news.ycombinator.com/item?id=660986

Not to impugn on your care packages, but if your friend wants such more often, there's a store in San Jose that sells them:

http://www.aussieproducts.com/contactus.asp

(They do mostly mail order it seems, but they have a storefront in SJ)

It's interesting Kogi BBQ is mentioned, because that to me is an example of the dysfunction around Twitter.

Twitter started out as an SMS service, yet these days they cannot reliably deliver SMS messages. Kogi is awesome food, so whenever I visit LA, I turn on SMS delivery for Kogi's messages on Twitter. This would be a really great feature, since when I'm out for the evening I may try to stop by Kogi for some food, and it's great to get updates about where they are in relation to me, and if they're running late due to traffic (which happens a lot in LA). Except the last two times I was in LA, Twitter didn't send about half the updates to my phone. Which makes it not useful at all.

The Kogi case is a great application for one-to-many SMS updates, but Twitter completely drops the ball on that. Yet Twitter is so popular, there isn't really room for a more technically competent service to provide value. As we've seen already, merely cloning Twitter, even if you do it better, isn't enough.

The world could use a popular, reliable one-to-many SMS delivery platform. Instead we're stuck with Twitter. It's sad.

> Even the editor of Bon Appetit wrote about it, How many other food trucks do you think she writes about? Major PR score!

That's probably more due to the fact that Kogi made the New York Times and the LA Times, than due to direct Twitter awareness. Kogi often has long lines now, but that pretty much started happening when it hit traditional media. The first time I went to Kogi was last December, before it hits NYTimes/LATimes, and there was no line.

Definitely Twitter helped them hit their early following, but it doesn't seem that a majority of their customers now use Twitter. Last few times I've been waiting in line for Kogi, most people I talked to around me got the word that the truck was there due to their website itself, or word of mouth from friends.

> I didn't follow balloon boy but I can picture the scenario that anyone interested could follow the balloon's path in real-time

You could have also followed it on CNN. It's funny that Balloon Boy is an example. That story was not broken on Twitter, but via traditional news channels, and all the Twitter chatter was based around updates from those same traditional news channels. Except if you tried to follow the story on Twitter, you'd have way more noise to filter through, as Twitter has a lot of redundant posts (Retweets don't add any value when you're searching).

> Yep with -arch i386 it appears to work but the profiling info it returns is all zeroed.

That's the story of OS X since from when it was first released. For example, you'd have some functionality advertised as being in the system libraries (like, oh, certain important pthread functions) so they'd pass the configure checks, but they would be in reality just stub functions that mean nothing. So your program would subtly fail.

That particular problem was fixed in 10.1, but it seems like every release has a few things like that: something from UNIX that looks like it's there, but actually amounts to no-ops. It's maddening.

Traffic in SOMA is a lot less around evening rush hour, because there isn't a mess of cars backed up on the streets leading up to the I-80 enterance.

Traffic is worse, however, around the other bridges in the area, since the interchanges to the major freeways (101, 880) at those bridges do not have the capacity to handle the offload from the bay bridge traffic. Some of those interchanges only have one lane vs. multiple for the bay bridge entry and exit points, and not enough merge distance to make things smooth.

Parking at most BART stations is also pretty much at capacity, whereas usually there is plenty of empty spaces.

http://en.wikipedia.org/wiki/Eldfell

5,300 had to be evacuated from the island of Heimaey due to an eruption in 1973.

Huh? Plenty of things needed more than 4 GB RAM in the 1990s. Scientific and database applications, to name a couple examples. That's why IBM, Sun, HP, SGI, DEC, etc. were all shipping 64-bit machines then. There were also a lot of contortions implemented to make >4 GB happen on 32-bit Intel machines too.

What AMD did (and Intel subsequently copied) with x86-64 was to significantly drop the price of 64-bit capable hardware, to the point where it shipped on personal workstations which didn't really need it then.

"Only wimps use tape backup: real men just upload their important stuff on ftp, and let the rest of the world mirror it."

-- Linus Torvalds

I thought it was a shoplifting deterrent. SD cards and the like have bulky, hard to open packaging to make it harder to conceal when walking out of a store. Amazon doesn't need such deterrents, so I'm glad their finally using their clout to get easier to open packaging.

I agree that worrying about leaking info to competitors is silly, but...

> If they aren't good enough for your team and your competitor ends up hiring them, what does that say about your competitor?

There are a number of reasons you may say no to a candidate who is highly qualified. They may not be a cultural fit for your specific organization, or you may have more than one qualified candidate but only one position to fill. So just because you wind up rejecting them, doesn't mean they suck.

I disagree. Writing multiplatform makefiles without explicit conditionals is cumbersome and ugly, mainly because the problem space contains so many exceptions. In this case, purism isn't really practical.