Friday, April 27, 2007
Bill Moyers reports on the behavior of the media as it helped the administration make its case for war in Iraq in 2002. It's about an hour and a half to watch it online, but it is riveting if you have the time.
Sunday, April 22, 2007
I like to say that the this topic is partly inspired by Anna Nicole Smith, since around the time I thought of it, Smith died and for some reason completely monopolized cable news for several weeks. I kept wondering: Why in the world do they think people care about her? People die all the time. As celebrities go, she wasn't particularly interesting. Do people actually read this stuff?
Web 2.0 can give sort of a handle on answering this question. At Digg.com and similar sites, people actually rate the news by voting it up or down. A given news item will get an overall "score" for how many people voted for and against it.
Now suppose you take the average rating of a news story on a given subject -- let's stick with Anna Nicole Smith as the example -- and compare it to the number of times that that subject story appeared in the news, across all news sites. The first number would tell you what people want to read about. The second number would tell you what is being presented most often as news. We could probably normalize this by what section of the newspaper it appears in -- for example, a story that appears on the front page is considered more important than one that doesn't; a long story may be more important than a short one.
So the question at hand is: how successful are news sources at generating information that people want? Are readers really treating their news as entertainment, or do they recommend hard hitting investigative reporters much more heavily? And what about media bias, either liberal or conservative?
In theory, it may be possible to quickly identify stories as leaning towards a liberal or conservative position, perhaps by cross-referencing them with the people who recommend them. Then what? Well, suppose it turns out that there are more liberal stories than conservative ones in the media... but suppose also that the liberal stories tend to be rated higher and read by more people than the conservative ones. That might indicate that, for instance, the idea of what "liberal" means is out of sync with the political center. Of course, it could go either way, and I'll be interested to try to come up with a measurement that doesn't bias the results.
There are tons of flaws with this topic, and I'll acknowledge some of them up front. For starters, those who subscribe to Digg almost certainly do not constitute a representative sample of all people in the country who read the news. So there's no way I can think of to justify any claims about all people nationwide. However, just investigating this cross section of people, and seeing what they like, could be useful and interesting in various ways that I haven't thought of yet.
When I talked about this topic with Dr. Ghosh, who will be my adviser, he said I shouldn't get sidetracked by that kind of problem, because it's not unusual for a research paper to be limited in scope. In fact, he recommended that I deliberately limit the scope to around five news sources, so that I have interesting things to say about just articles from those sites. I was thinking of picking three somewhat "mainstream" media sites (for example, NY Times, Washington Post, and CNN); then pick a liberal feed (perhaps Daily Kos) and a conservative feed (Fox News? Washington Times? WorldNet Daily?) to compare against.
Wednesday, April 18, 2007
This seems like as good a time as any to dredge up the term paper that I did with Chip Killmar for Data Mining last year. I'm at least somewhat familiar with the politics of gun control, although it's one of the positions that I doggedly refuse to take a firm position on it.
Many people cite statistics that claim to show that states which allow concealed weapons have less violent crime than states that do not. In fact, there are a lot more factors which contribute to violent crime rates. The most often cited expert in favor of concealed-carry laws is John Lott, who has written several books under titles like More Guns, Less Crime. However, Lott's methods are extremely suspect and generally not very convincing, for reasons we go into in the paper.
Density of population is the leading contributor to overall violence levels. States with large, crowded cities tend to have much higher populations than those with mostly small, rural areas. Not (necessarily) coincidentally, small rural areas are much more likely to have a strong NRA presence, and those are the states that tend to pass Shall-Issue Right-To-Carry laws, giving nearly all citizens easy access to permits that allow them to pack a concealed weapon, barring criminal records and other extreme circumstances.
As a result, people who claim that gun-friendly laws are successful often point to raw crime statistics, correctly stating that states with RTC laws have less gun violence on average, but failing to note the other significant factors such as density.
Also, throughout the 1990's, crime decreased nationwide. Since previous studies of the effectiveness of RTC laws mostly occurred during the 90's, they showed crime decreasing AFTER the passage of these laws, and invoked the common "post hoc ergo propter hoc" fallacy to say that the laws themselves were the cause of reduced crime. Yet crime also went down in areas that didn't institute such laws, often (though not always) even more than in pro-gun states.
Currently 37 of the 50 states have shall-issue laws, with more than half having been instituted after 1987, when Florida become a leader in doing so. Our project compared cities across the US and factored out the population size of each city. Anecdotally, some of the highest crime rates in the country are found in cities like Miami, Dallas, and San Antonio: all of which are very large cities which happen to exist in states (Florida and Texas) with long-standing RTC laws. By contrast, New York City, which is frequently associated with high crime, actually has some of the lowest violent crime of any major city, and New York State remains one of the states that has not issued any law favoring concealed weapons.
Our final results were inconclusive, because the error bars are very high and it's difficult to do a reliable city-by-city comparison without knowing more about the contributing factors. However, the data tentatively indicates that there is indeed a somewhat significant increase (mostly around 10-20%) of murder and rape in the few years following the institution of a RTC law, over a city of similar size which did not issue that law over the next several years.
By the way, Virginia has a shall-issue law. They have since 1995. In many ways the state of Virginia is a shining example of what kind of laws that the NRA would like to see instituted nationwide. I've read that even in RTC states, very few people actually carry concealed weapons. So often the theory is floated that in a state with an RTC law, crime is deterred because criminals are SCARED that they'll get shot. Clearly this was not the case yesterday, nor is it surprising that a lunatic going on a shooting rampage is unlikely to consider the finer points of state laws.
Am I in favor of banning guns? No. Even if it were conclusively proved that a certain level of crime were caused by guns, I think the case can be made that constitutional principles override an outright ban. After all, even freedom of speech and the press certainly causes harm, but we believe that the principles of free speech and press are often more important than a little additional safety. On the other hand, I don't see any serious problem with making people jump through more hoops than they currently do before they can use a gun, just as we make people get driver's licenses and it's possible to revoke those licenses.
Could a pistol-packing student have prevented the massacre in Virginia? Nothing's impossible, but consider that this was an extremely rare event. In order to create a happier ending, a heroic student would have needed to be present, who happened to be packing that day, had the presence of mind to shoot the guy, didn't get himself shot first, and didn't make the situation worse by shooting innocent bystanders in the process.
Furthermore, suppose that there are TWO heroic students, who each hear the shot and whip out their concealed weapons. They don't know who's firing, but they see each other holding a loaded firearm. How do we know they don't shoot each other? When the police arrive, how do they know that they are not accomplices to the crime, and gun them down? Now multiply this risk by the number of days that numerous students are walking around campus with loaded guns, and a particularly crazy guy is NOT walking around campus (i.e., almost every day, in every university). Are we really saying that all these extra guns create no additional opportunities for more incidents?
I don't know the answer to that, of course, but neither does the NRA. That's why I'm mostly for people's right to own as many personal firearms as they wish in their houses, preferably with some kind of mandatory training; but I'm still extremely dubious about these concealed weapon laws.
Saturday, April 14, 2007
- Go out to dinner with Ginny and Ben.
- Come home.
- Bed at around, oh, 7:30.
Sandy, sitting in front of me, is browsing shoe sales online. So I'm not the only one who is using the internet to escape paying attention to detailed explanations of the syntax of RTCTL. I'll pick it up later in the lecture slides and future study groups, at least I hope I will.
Next month, worse than two months ago, I am responsible for TWO homeworks and TWO tests. The good news is that unlike two months ago, I have four weeks to prepare instead of three; the Requirements homework is meant to be easy, and the Real-Time Systems homework involves playing with a computer program, something I'm pretty good at.
I haven't been excited about my classes this semester, but I seem to be doing well in them based on a slew of returned assignments where I beat the class average. I may get some more A's under my belt. Next semester I'll be taking a summer topic on Web Server programming. There's something I should have learned a long time ago.
In other news, I spoke with Dr. Ghosh (my old data mining professor) today, and he likes my idea for a Master's Thesis. I will soon post an update to Operation: Help me with my thesis. I want to thank everybody who contributed ideas in the comments; your feedback was very valuable and helped me come up with the germ of a topic. It needs a lot of fleshing out still, but Ghosh is sufficiently interested to be my adviser, and he told me he'd put me in touch with some of his former students who work at Yahoo and know how to do the kind of text-spidering that I'm going to need to start doing in the coming few months. More details later. In any case, it can't hurt to have contacts at Yahoo, since this is a topic of interest to me.
Funny story about lunch today. We get an hour between classes for lunch time. A group of people decided to head for a new California Pizza Kitchen that had just opened. Well, that was a mistake. The place was packed and slow. We didn't manage to leave until ten minutes after class had started. We got our pizzas to go, but one person (not in my class) grabbed the bag and took all the pizzas. I met him during the first break, but it was after 2:00 before I got to enjoy my barbecue chicken pizza, at which point it was lukewarm. Still pretty good though.
Yawn. Still going to be a long two hours. Okay, Dr. Mok says that the rank of several nodes in this graph is infinity, because you rank it by the maximum path length and you have the option of going into an infinite loop. Yeah, whatever.
Thursday, April 12, 2007
It's not a very frugal choice compared to, say, Mountain Dew. But it is both cheaper and easier than actually going to Starbucks or Seattle's Best down the street and buying something from them.
Of course, as everyone knows, Starbucks is evil. I guess I should start feeling guilty now.
Monday, April 09, 2007
My sister sent this video to me. I rarely post plain old silly stuff here, but I thought this performance was really outstanding, and I appreciate a guy who uses his obvious musical talents for comedy purposes.
Also, Ben is a big fan of the Weird Al version ("Bohemian Polka"), which is by far his most requested song, so of course he loves this video too.
Saturday, April 07, 2007
Usually when anyone complains about government observation of religion, they are accused of persecuting Christians by preventing them from freely exercising their own religion. ... But as we can see, the same Christians who insist on their right to express themselves are not willing to afford the same "right" to Muslims, and that's the point where they fall back on declaring that only Christian prayers have a place at the table.
Go read the rest here.
Thursday, April 05, 2007
Now that George Bush is a lame duck, oh who will be brave enough to come out and tell us how great things are going in Iraq? Answer: John McCain to the rescue!
Last week, McCain goes on Wolf Blitzer's show and snidely admonishes Wolf for suggesting that things are not going so well over there. McCain lectures Wolf, saying:
"You know, that's why you ought to catch up on things, Wolf. General Petraeus goes out there almost every day in an unarmed Humvee. You want to -- I think you ought to catch up. You see, you are giving the old line of three months ago. I understand it. We certainly don't get it through the filter of some of the media.
Later in that same show, Baghdad correspondent Michael Ware responds:
"No way on earth can a westerner, particularly an American, stroll any street of this capital of more than five million people.
I mean, if al Qaeda doesn't get wind of you, or if one of the Sunni insurgent groups don't descend upon you, or if someone doesn't tip off a Shia militia, then the nearest criminal gang is just going to see dollar signs and scoop you up. Honestly, Wolf, you'd barely last 20 minutes out there.
I don't know what part of Neverland Senator McCain is talking about when he says we can go strolling in Baghdad."
You can watch the video of Ware berating McCain, it's great fun. Michael Ware sounds a bit like the late Steve Irwin, and Possum Momma mentioned that she kept expecting him to yell "CRIKEY!" and leap on top of McCain. Except, of course, that they were half a world apart at the time.
Anyway, the comedy could have just ended there. But obviously McCain never learned the rule that you always leave the audience wanting more. So what did he do? He gave us more.
Earlier this week, McCain and fellow Republican Senator Lindsay Graham went on a field trip to Neverland. As the intrepid Senators proved, it's perfectly safe for a couple of high ranking members of the U.S. Government to walk around freely in a perfectly ordinary Iraqi marketplace. Graham later gushed: "We went to the market and were just really warmly welcomed. I bought five rugs for five bucks. And people were engaging."
So as you can see, it is easy to walk around unharmed... as long as everyone involved is wearing bulletproof vests.
surrounded by 100 American soldiers as bodyguards.
escorted by 3 Blackhawk helicopters.
OH YEAH, AND ALSO...
2 Apache gunships.
Yep, perfectly safe. At least until after all those troops left. The next day, the marketplace was bombed, killing 14.
By the way, if you're looking for an ideal hot spot for your next vacation, why not consider a visit to scenic Baghdad?