Marco

From this NYT article about very severe recently discovered flaws in Diebold voting machines:

David Bear, a spokesman for Diebold Election Systems, said the potential risk existed because the company's technicians had intentionally built the machines in such a way that election officials would be able to update their systems in years ahead.

"For there to be a problem here, you're basically assuming a premise where you have some evil and nefarious election officials who would sneak in and introduce a piece of software," he said. "I don't believe these evil elections people exist."

What planet is that guy from? More importantly, why the hell is our democracy being entrusted to his company?!?!

You knew it was coming...the RIAA and MPAA finally joined forces.

The U.S. Supreme Court decided the MGM vs. Grokster case today in favor of MGM. So what does this mean?

My first thought was that this is A Very Bad Thing. This case has been in the works for quite some time now, and I've always sided with the geeks against the content cartel. But it's interesting to take a look at the actual court opinion. I read most of it, and I was both surprised by some of the facts I learned and hit by how reasonable the opinion is. (I haven't yet read the concurring opinions of Ginsburg or Breyer.)

The main argument in favor of Grokster that has stuck in my mind the last few months is that the 1984 Sony Betamax case should apply here because the tool (software in this case, hardware in the Betamax case) has "substantial non-infringing uses" and the company that merely makes the tool can't be held liable for the uses it is put to by the various and sundry characters who use it.

The problem is that the Betamax case isn't really relevant here. There are some nit-picking legal reasons why not that are described in the opinion, but the one that stands out for me is the following: it's clear that Grokster and StreamCast (which distributes Morpheus) are NOT merely the makers of an impartial tool that could be used for evil, but rather they actively promoted, encouraged, and sought the illegal distribution of copyrighted works using their software. They actively courted former Napster users, promising them both implicitly and rather explicitly that the copyrighted works they found on Napster would abound on Grokster/Morpheus. They responded to emails from users asking how to acquire and play copyrighted material, helping those users in quite direct ways. And they also had plenty of internal communication that shows they saw distribution of copyrighted materials as the core use of their software. StreamCast's CTO even said at one point "[t]he goal is to get in trouble with the law and get sued. It's the best way to get in the new[s]."

If you're wondering about this case and its result, I encourage you to read the opinion. It was enlightening for me--I realized this wasn't just a case about openness and technology vs. stifling innovation. If you think that mass distribution of copyrighted songs/movies isn't or shouldn't be illegal, that's one thing, but short of that it seems to me that the Supreme Court hit the nail on the head when it said that Grokster and Streamcast are in the wrong.

What we need now is for a truly impartial filesharing system to become popular and have more substantial actual non-infringing use...

I was catching up on some nonplatonic reading when I read this post by Graham (and its comments) about Starbucks coffee sizes. They bug me. Graham likes them. Therefore, I rant.

Since I don't often go to Starbucks, whenever I do actually go (or when I encounter their nomenclature at another coffee shop, which is unfortunately happening more and more), I have to study the menu for a minute to remind myself whether "tall" is bigger or smaller than "grande" and so on. Once I see the order of the sizes on the menu that lets me translate into small, medium, and large, I can order my drink. But whenever I say "tall" or "grande," it feels like I'm capitulating to a trend that's both smugly corporate and painfully trendy.

And contrary to what Graham and Ben say, it *does* matter what size a "venti" is--since venti is Italian for 20, as a 20 oz size it actually makes the most sense of all the names!

Enough for now...maybe I'll read nonplatonic again in another two weeks and see if anyone (Graham?) has protested.

My computer rebooted last night. Twice. Once at 11:24pm and again at 2:00am, though now its clock is an hour behind, so who knows when it really was.

I can't figure out what happened--I don't think the power went out, since my alarm clock wasn't reset. Did my computer get hacked into? Maybe, though the only port open is 22 and I haven't heard of any recent ssh vulnerabilities (and although I do regularly get batches of ssh login attempts for common usernames and/or root from far-flung IP addresses, I seriously doubt that's it).

My system logs aren't particularly helpful--nothing shows up right before the reboots.

Annoying.

I was at lunch with Jessica today in a nice little Japanese place near campus. I overheard two people at a nearby table saying "reduction in rank" and "kernel" and thought they must be from the statistics or CS department, and was curious exactly what they were talking about.

Then I heard them say something about "punishment in the military."

Yesterday Ecuador's president, Lucio Gutierrez, dissolved the supreme court and declared a state of emergency. Today he lifted the state of emergency, though he now faces even more calls for his resignation.

Ecuador hasn't had good luck with its presidents...there was Abdala Bucaram, not-so-affectionately known as "el loco," who has just returned from 8 years of exile after being president for a short time. Gutierrez was elected with wide support among the population, but he's fallen from favor due to his dealings with the IMF that have been very non-beneficial for Ecuador's people.

The overall rampant corruption in Ecuador doesn't help things either.

Collin wrote:

I'm not saying that there are huge error bars on something like the infant mortality rate in a country like the US, but they're still there. Basically my point is that there will allways be sample error.

Right--unless you're using a population instead of a sample. Take birth rate, for instance. What I was trying to say before is two things. First, if you do actually measure all births for any given year, then that number is an actual count, not a sample, so there is no sample error, so you would need no error bars. Second, it seems completely infeasible to get all births on record because there are just too many people who are under the radar of the system. No I wasn't trying to say that the U.S. government is aware of every birth; I pointed out that the very poor and the illegal immigrants are likely not to be counted, and others are sure to fall through the cracks as well.

Ben wrote:

I think the best way to model it is to treat the population as drawn uniformly from an infinite distribution since drawing a full population from a finite distribution would conglomerate additional stochastic processes with example selection.

Ben, I don't think I understand exactly what you're saying. Let me explain the way I understand it, and maybe you can respond if you understand things differently or if I'm completely missing what you're getting at.

When you put "error bars" on graphs, what you're doing is showing confidence intervals, typically 95% confidence intervals, meaning that you're 95% certain the true value lies within the interval shown. This interval can be calculated from the standard error, or sample standard deviation, given by the formula to the right. (The 95% confidence interval is approximately the estimated mean plus or minus twice the standard error.) Note that this depends on the sample size n but not the population size. We can use this statistic because of the assumption that the population is sufficiently large compared to the sample size that drawing without replacement approximates drawing with replacement. It's also based on the assumption that the samples are drawn uniformly at random.

Both these assumptions fail to hold in the case of looking at all recorded births for 2004 in the U.S. First, the "sample" you're looking at is pretty close to the size of the population (an extremely pessimistic estimate is one order of magnitude off, which is still too big for this assumption to work). And second, you're getting a very biased sample if you're recording people who are born in hospitals and/or who report births and not recording people who are born at home or somewhere else out from under the government's eye. So if the numbers from the CIA world factbook were calculated from "complete" birth records and so on, then I suspect that they made the assumption that the records truly are complete and presented them without confidence intervals since they are true numbers. Or would be if the records were truly complete and accurate. In any case, I think you'd need a pretty sophisticated model of how many people are born without being recorded in order to say anything more definitive.

Of course, you could take an alternate approach. You could take random samples from the population and find out how many births there were in your sample, then use that to estimate the rate in the population (complete with error bars to account for sampling error!). If you thought that you could take these random samples in a manner more truly representative than the difference between births reported and births unreported, you might get a better estimate this way. But you'd probably be going on one of geography, driver's license records, voter registration, or phone book listings, and these probably wouldn't get you results significantly more representative than birth records.

Maybe that's what you were trying to say, Ben. Anyway, I think there was something else I meant to write but this is too long anyway, so I'll stop here.

Rice said she believed strongly in the "role of debate, the role of the open and free exchange of ideas."

But, she said, when decisions are taken, "I fully expect that people will support those decisions because there is only one president of the United States and that's President Bush."

I think I like it, though it still probably needs a bit of polishing.

This text has taken the path English -> Greek -> French -> Dutch -> English:

We gardons this truths to be clear, which all persons he right is created, qu they are financed by their author thanks to some unalienable rights, which are these life, freedom and the search of the luck enters. That for for they the these rights guarantee, the governments between the persons have been determined, the authorisation controlled by their strengths of exactly that when n sails in which form of government the destroying of it extrêmes becomes, it draw human right for changing are or qu withdraw, and qu the new government establish, who is foundations to such principles lays and which is strengths under their such form organises, if maggots will seem as of now probable for influence the security and their luck.

You'll probably recognize it, mutilated though it be. I think it's great...especially the last phrase. Hooray for babelfish!

Apparently, in Malaysia some car thieves cut off a guy's finger to use it to start his fingerprint-protected Mercedes.

Now the only question is how the finger got to San Jose from Malaysia...

A week ago, my laptop's hard drive croaked. Ironically, it happened while I was in the middle of backing up my data from it to my new desktop. I didn't get all the data copied, but I did get most of it. That was the first time I've backed up the data on that computer in a year or two, so it would be fair to say I'm a lucky bastard.

Today I finally got around to submitting a warranty service request to IBM (it's under a 3-year warranty that expires this June). I had to register with their service website, and then I submitted the service request at 9:35 am.

At 10:10 am, I received a call. It went something like this:

IBM: "Hi, I'm so-and-so with the IBM repair center. Is this Marco?"

Me: "Yes, this is Marco."

IBM: "I have a repair ticket here for a laptop hard drive. What is the FRU number on the drive?"

Me: "Sorry, I don't have the computer in front of me."

IBM: "What's the size of the drive?"

Me: "20 GB."

IBM: "Okay, I'll have a replacement drive sent out to you. It should be there in two business days."

Me: "Wow, that was easy. Thank you."

Amazing. All tech support should be that easy.

Just a quick first post to try this out...thanks, Ben, for setting it up.