Robot Monkeys

So CNN publishes an article about the Great Tomato Scare of 2008 (If anyone wants to use that, you owe me a buck.), and the headline completely misrepresents the article. The headline: “Health Officials Question Whether Tomatoes Behind Outbreak“. And what’s the quote from the story hightlights?
“FDA: ‘There is a strong epidemiological association with tomatoes.’” So where’s the source of the fear?

“Produce investigations are very difficult, because a lot of times, vegetables are eaten all together,” said Dr. Patricia Griffin, chief of the Enteric Diseases Epidemiology Branch at the Centers for Disease Control and Prevention in Atlanta, Georgia. That makes it hard to trace back any one item to a source of contamination, she added. “We continue to keep an open mind about the possible source of this outbreak, as does FDA.”

Dr. David Acheson, associate commissioner for foods at the Food and Drug Administration, agreed. “There is a strong epidemiological association with tomatoes,” but the agency is also “looking into other ingredients,” he said.

So you have the scientists basically saying, “While we can’t be 100% sure it’s not from the tomatoes, we really really think it is.” Which of course means the take away is, “Scientists Dumbfounded!”

Now Occam’s Razor and Hanlon’s Razor says that I should attribute this to pure laziness, and that’s probably true. After all, the one story from Billy’s journalism classes is the time the TA or prof tried to explain the concept of a quorum using the example of “one more after the half.” Apparently this was too much more most of the students who would calculate the number required for a quorum as simply half, or one less than half. At least one proclaiming, “We’re not math majors!”

The future of American journalism at its finest.

However, I can’t help but think that this CNN headline is more indicative of American attitudes towards science. Nothing is known for certain, so nothing is known, and therefore everything is equally likely.
I believe this attitude comes from the poor state of math and science education in this country, and the need for “fair and balanced” coverage.

I want nice RSS reader. I’ve wanted one for a while now. I even made one. It’s pretty cool, for being a ticker. Only there’s one catch. It’s a GNOME applet.

Fuck.

You can’t install a GNOME applet locally. You have to install it in /usr, which makes it absolutely useless unless you’re the sysadmin. God damn it. There’s just no good reason to do that. That’s just laziness. Unix has always been (mostly) agnostic, and then here comes GNOME botching things up as usual.

I really don’t want to have to use something Tickershock, since I already have something better. Well maybe I’ll give in.

Linux, can still blow me

Bank of America’s website lets you download statements in a variety of formats, including Quicken and “Printable text” and “Excel”. However, the “Excel” format isn’t Excel at all. It’s CSV. And just to make things just a bit more ironic, Excel 2004 for the Mac (aka Excel 11) won’t open file named stmt.csv. (Yes, that’s really the name of the file.)

Uhura comes in Red Rock with a tethered hands free while carrying the phone in his hand.

Good thing he splurged.

So Microsoft is colluding with General Electric (via NBC-Universal) to break television and to move us 30 years into the past. That’s right. Microsoft conspired to remove the ability to record television shows. Why? They probably got paid off. I suspect this is part of the .

Should anyone really be surprised that this happened? Of course not! Microsoft shoved their ironically named, “Play for Sure” digital restrictions management scheme overboard and screwed all the those dumb enough to exchange money for something they don’t control. (This goes for both the end users and the device “partners.”)

So just in case any of you still don’t understand how things like this work. You are not a customer. You are a commodity. You don’t even have a seat at the table. Microsoft, GE, none of the megacorps care about you. You’re a resource to be exploited, and are bought and sold like a jailhouse bitch.

Yesterday I spent several hours trying figure out why g++ couldn’t resolve a set of nested templates. I figured it was a strange problem with how C++ worked. It didn’t make any sense to me because I could see how the templates should be resolved. I just couldn’t figure out why it wasn’t working. It certainly looked like inheritance was screwing up the template resolution. (This is the code.) I emailed John the code, and he didn’t have any problems compiling it. That’s when I realized, it’s a fucking compiler bug in g++ 4.1.2.

fuckyou-cpp.cpp: In function ‘void serialize(std::ofstream&, std::vector >*) [with V = std::basic_string, std::allocator >]’:
fuckyou-cpp.cpp:100: instantiated from here
fuckyou-cpp.cpp:18: error: no matching function for call to ‘serialize(std::basic_ofstream >&, std::basic_string, std::allocator >*)’
fuckyou-cpp.cpp: In function ‘void unserialize(std::ifstream&, std::vector >*) [with V = std::basic_string, std::allocator >]’:
fuckyou-cpp.cpp:114: instantiated from here
fuckyou-cpp.cpp:38: error: no matching function for call to ‘unserialize(std::basic_ifstream >&, std::basic_string, std::allocator >*)’

Update: Sat May 10 19:04:30 PDT 2008
So after much tribulations, I thought I figured out how g++ resolved templates. In fact, I thought I figured it out, right up until the moment I tested my theory. Then it worked. Then I intentionally tried to break the code, and it still worked. Damn it. I thought I learned something, only for it turn out that I didn’t learn anything. I learned all about -frepo (and how it can cause an infinite loop in some versions of g++), #pragma implementation and #pragma declaration, and -fno-implicit-templates (and how it works on the entire file, thus breaking STL code), and even different template resolution methods.

Now the code I wrote works without any of those switches, and without shoving explicit declarations in a separate file. I have no idea why. I asked Arek to test it on a newer version of g++, and he said that my fuckyou-cpp.cpp didn’t work on anything newer. But he did mention that he got it to work by just flipping the order of template declarations.

So I guess these are the lessons:

1. Make sure you put the vaguest template declaration first.
2. If you need an untemplated version (i.e. the template is fully specified), then put that in a cpp file separate from the hpp file that will be included in the other files. If you don’t you’ll get multiple declarations of the untemplated version since it’s now contained in multiple files.

Dear Madison Avenue,

Stop using the word “literally” to introduce a metaphor.

The Geico gecko says, “They literally hand you an english muffin with butter and jam. – That’s a complete a dramatization…” Some annoying woman in a vehix.com ad says, “You can go on the website and literally take a test drive. – You’re practically sitting in the car.”

“Literal” means

actual: being or reflecting the essential or genuine character of something; “her actual motive”; “a literal solitude like a desert”- G.K. …

without interpretation or embellishment; “a literal depiction of the scene before him”

limited to the explicit meaning of a word or text; “a literal translation”

avoiding embellishment or exaggeration (used for emphasis); “it’s the literal truth”

None of these definitions allows you to use “literal” to embellish the story. In fact, it’s literally forbidden.

Please stop. Please please please stop.

Honestly. Who taught you how to do this stuff? Alanis Morissette?

I’ve noticed that even with Microsoft’s advertisement infrastructure, Facebook has the shittiest ads. Like absolutely HORRIBLE ads. There’s no excuse for that. Facebook knows EVERYTHING to target ads to people. They know where you are. They know what you like. They know all your demographic information. They know all your friends, and everything about them. They have everything they need to do microtargeting of ads. Yet, they epically fail.

“Punch the monkey and win a free ipod”? That’s your ad? Geez.

Recently they’ve been serving “Date hot chicks near you!” to me.

Until today…

As far as I can tell, It’s because I set my status update to “Jonathan is a sexxy dude.”

Two billion dollars my ass.

Citibank’s AccountOnline.com’s login page is absolutely broken. If you have no idea what your “security word” is and ask for a hint, you’re essentially told that it could be anything, including nothing. That really isn’t much a hint, because its safe to assume that your security word, if it exists, expresses some concept in the universe of concepts. If you’re asking for a hint, you what concept you’re supposed to guess from. But no. Citibank tells you “It’s a valid string of characters, including the null string.” Gee. Thanks.

However, if it’s actually nothing, you’re screwed, because the security word is a required field, and so you’re not allowed to enter if your security word if it’s actually the null string, which they specifically say is a valid security word.

God! How the hell does this pass testing? Sadly I know the answer. There is no testing.

Dear Interweb,

Who do all the delegate counters suck?

I have found four different counters with four different answers. Some ignore superdelegates, which is dumb, since their votes will matter this year. But the pledged counts are wrong! God!

As I write this at 3:31am PST, we these are the counts

MSNBC:

Wikipedia:

CNN:

ABC:

DemConWatch:

Now DemConWatch also gives counts that include Michigan and Florida, just in case Clintonistas pull off their coup, which I like. But who knows what’s up with the the “Unknown Iowan Delgates”, that everyone else says are Edwards’s.

CBS:

AP:

Does this mean that we’re going to need some sort of metacounter that averages all the predictions?

Why is this so hard?