July 31, 2006


Holding no brief for Starbucks I an nevertheless compelled to respond to this pablum in The New Republic:

About fifty years later, there are dozens of coffeehouses in every major city in the United Statesbout fifty years later, there are dozens of coffeehouses in every major city in the United States .... They are all called Starbucks .... Starbucks is a state for our day, a commercial society organized within psychographic, rather than geographic, borders--parameters that are now more meaningful than the old rivers, mountain ranges, roads, and lines of longitude and latitude that cable TV, the Internet, and cell phones render moot .... The Starbucks habitu├ęs are united [by ... ] the aspiration to belong to the young, white, moneyed community that we all perceive Starbucks to be .... As a class, Starbucks's various compilation CDs are the audio equivalent of those trade paperbacks that impart the hundred and one things a person needs to know to be culturally literate.

Where does The New Republic find these people to write such nonsense? The independent coffee shop has hardly been supplanted by Starbucks, complemented perhaps. A state for our day? Isn't there just a bit of a difference between a chain of commercial outlets for a nutritionally neutral, if mildly addictive and enjoyable, beverage that millions of other people manage to do without; and the nation state that compels our behavior through taxes, laws and violence? And the clientele? Isn't this going to be highly dependent on which branches of thousands the writer visited?

Cultural criticism meeting the mass market is never a pretty sight.

Posted by eroberts at 2:32 PM

July 29, 2006

Sculpt what ...

An email arrived from Runner's World (really, I think the apostrophe should be plural, but whatever) today offering me an annual subscription that would probably equal the amount I spend on the magazine in airports over the course of the year. But frankly I refuse to pay money for a magazine that purports to be about running but opens their sales pitch with the line I have highlighted below.

I will say this though. The daily interviews on the website are [sometimes a little New York centric] mostly pretty good, and I would consider ponying up the money if they charged separately for that. I heard someone muse once that the overlap between the magazine and the website readership must be really low. Do the editors know that? Do they not care? Are they happy to subsidize their coverage of the competitive side of the sport off the subscriptions of the more risk-averse side?

Posted by eroberts at 1:02 PM

July 27, 2006

Drugs on ice

Exhibit A: Tour de France winner Floyd Landis failed a drug test. Shooting yourself up with testosterone. Bad.
Exhibit B: World Anti Doping Agency considers banning rooms that simulate sleeping at altitude. Debatable.
Exhibit C: Paula Radcliffe swears by an ice bath after every race. Good.

Drugs don't make you faster. Even if I riddled my body with needle marks and someone else's EPO enhanced blood I'd be lucky to ever break 15 minutes for 5000m. Drugs do help you recover quicker from hard workouts though. So do icebaths. If you haven't discovered the icebath and are training for a marathon, really, now is the time to start before they ban them too. In fact, I'd say that sometimes the icebath after the hard workouts seems to be the difference between 2 and 3 workouts in a week, because it seems to cut my recovery time by up to 12 hours.

I don't think they're going to ban icebaths anytime soon, but my point is that where you draw the line on unacceptable aids to athletic performance is not clear. The rationale that drugs are bad because they may cause harm to athletes is one basis for discriminating between drugs and artificially simulated altitude, but even here it's a matter of probabilities and proportions, not absolutes. Jack your altitude settings all the way up to 30,000 feet and you'll cause some harm to the athlete. Stay too long in an ice bath that's too cold and you could kill yourself. It's harder to do than overdosing on drugs, but still just possible. And once you're dead, what does it matter?

This is to say nothing of the probably even greyer area that you reach when athletes have legitimate medical conditions that may require drugs that also have recovery-enhancing benefits. Or, the difference between "natural" and "synthesized" products. Getting protein, carbs and liquid within 15 minutes of a workout is also a significant boost to recovery. It's significant enough that you sure notice the difference when you don't replenish straight away, and then try to run twice that day, or do another moderate to hard day the next day.

No-one is ever going to be worried about what I'm doing in between workouts, but I'd wager it's a good deal easier for me to get in the icebath and the full meal after my morning run than someone with children and a job that expects them there promptly. You could extend this contrived example of inequity to more accomplished athletes, the sponsored runner who only works a 20 hour job because they have a shoe contract versus the person 20 seconds slower than them over 5km who doesn't quite have the sponsorship and works full time. But where do you stop with inequity as a basis for banning recovery-enhancing substances and techniques? We can't sponsor everyone, nor could we really level off the playing field by making everyone wait four hours after working out before getting anything to eat.

Drugs are bad, I agree with that. Icebaths are good [for you]. But in between them there's a lot of grey and the ethical basis on which we draw the line is not that clear.

Posted by eroberts at 1:00 PM

July 26, 2006

What will they think of next?

Rubber sidewalks. Made by this company. I guess you couldn't really call them a "rubber pavement" (pavement being the British-Australasian term for sidewalk) since pavement kind of implies asphalt or conrete or cobblestones (little known trivia for runners: most sidewalks in Australasia are, in fact, asphalt, not concrete, meaning you don't face the devils choice of running in the street for ease on your joints but have to look out for cars versus running on the sidewalk to save yourself from lunatics in cars with greater stress on the body).

Anyway, the rubber sidewalks. What a thing! Think of the possibilities! Could you paint them red and put lanes on them with markers every 10 metres for pace judgment? What about a little inside rail to stop people cutting the corner on the way to the store or school?

Posted by eroberts at 4:01 PM

July 24, 2006

Avoidance strategy

Last week's Minnesota Poll in the Star Tribune revealed something a little odd, but nevertheless predictable (and predicted): Tim Pawlenty is slightly ahead of Mike Hatch in the governor's race, and Amy Klobuchar has a big, big lead in the Senate race. Who are those Pawlenty-Klobuchar voters? My guess is they are largely found in the western and southern suburbs of the Twin Cities, where voters are familiar with Klobuchar from her role as Hennepin County attorney, less familiar with Mark Kennedy, and well-disposed towards our archetypically suburban governor, Tim Pawlenty.

So, I would like to reprise my campaign prediction from 2 months ago. The dynamic between the Senatorial and gubernatorial candidates in Minnesota is going to be interesting. Look for Tim Pawlenty to steer well clear of Mark Kennedy during the campaign, but look also for Mike Hatch to try and tag along with Amy Klobuchar. Although Pawlenty was slightly ahead of Hatch, there were two warning signs for the incumbent. Hatch had lower name recognition and higher favorability ratings, yet was still only just behind. If I had more time for this kind of thing I'd keep a running tally of joint campaign appearances by the top-of-the-ticket candidates ...

Posted by eroberts at 6:41 AM

July 23, 2006

Global warming

Doesn't global warming give the lie to that old phrase "there's nothing you can do about the weather"?

Posted by eroberts at 7:41 AM

July 20, 2006

Simultaneous marathon running

As if there weren't enough websites that aggregated race results and tried to advertise stuff (I won't link to them, if you care you know what I'm talking about) there's also Virturace. Now I'm well aware of why database programmers in this kind of area don't like to delete entries—it's tricky programming and you inevitably wipe out some legitimate entries as well as silly duplicates—but surely they should be able to not make it seem like I ran the Philadelphia marathon three times in one weekend in three different times last year.

For the record, the 2:49:04 is my official chip time, while the 3:07:46 was done by my alter-ego (who pronounces his name "Evv-in" ...) who lives in Bethesda, MD. I was sorry to miss out on the chance of meeting him last year at the Stonewall Jackson Ambulance Run.

Posted by eroberts at 10:29 AM

July 17, 2006

Two out of three ain't bad

and "If you have a man servant, take advantage of him" ...

With an overnight low of 82* Sunday's run had the potential to suck. Six hours sleep, a 5am starting time after waking up at 4:45, and 15mph headwinds the first half hour made me wonder why I was out there. The surprising thing was that I saw five people running the other way before 5:30am, so my strategy was shared by other crazy people. But it didn't suck. It was fun. If you have company for two of the three hours an easy 38km doesn't even feel that long. Thanks, Zeke.

As we were coming out of Minnehaha Park towards the end of the run we were passed by a young woman moving relatively quickly (7:30/mile) with apparent ease. She slowed and chatted to us for a while, and was training for Twin Cities. She had the assistance of her partner/boyfriend on a bike carrying her water. Literally. Every runner's dream. A potentially gigantic waste of time for the person on the bike. Even if the runner is doing 6:00/mile that's not much of a workout on the bike. Now, maybe this is just me, but if I had a mule beside me with water I'd keep running when I was drinking.

After the woman with the water-carrier had passed us, and Zeke had left me to head back to his car, I passed her a mile later as she was stopped to take a drink. I thought maybe they were sheltering from the storm that was passing through to drop the temperature from 81 to 73 over the course of the run (a much better temperature progression for 38km than the reverse!), but no. Glancing around I saw that having taken the drink she was off and running again. Odd. Projecting marathon times from long run pace is rife with inaccuracies, but if you're doing a two hour run at 7:30/mile you're probably looking at doing a marathon between 2:55 and 3:15. Decent running. Locally competitive for a woman runner. If you're that good it seems odd to be practicing stopping for your drinks.

Posted by eroberts at 7:17 AM

July 13, 2006

Writing hierarchical datasets in SAS

Really, I don't expect anyone among my regular readers [of my irregular writing] to be interested in this. But do say in comments if you are ...

I had cause today to have to write a hierarchical dataset. Now, reading a hierarchical dataset into a statistical program is so routine that the internet is rife with examples (SAS, Stata). Perhaps my Google skills are rusty, but not so much help from the internets with writing a hierarchical dataset. (Does anyone else have trouble typing hierarchical? I feel like I'm spelling it wrong at least half the time).

Anyway, I digress. The example code below assumes that you have a variable that uniquely identifies the unit that contains other observations. Or, more specifically you might have a household serial number or a family identification number.

SORT your data. The first step is to sort the data by the household or family serial number. If you have some identifier within the family or household (a person number, unique within the household or family, for example) it helps to sort by that too.

Then it's straightforward

DATA _NULL_ ; /* No need to write a SAS dataset at the same time */
SET <file you are reading from> ;
BY <household or family identification number> ;

FILE "<full path and name of the file you want to write>" ;

IF first.<household or family identification number> THEN DO ;
PUT <first household variable> <format> <begin-column> - <end-column>
<second household variable> <format> <begin-column> - <end-column>
<last household variable> <format> <begin-column> - <end-column> ;

PUT <first person variable> <format> <begin-column> - <end-column>
<second person variable> <format> <begin-column> - <end-column>
<last person variable> <format> <begin-column> - <end-column> ;

Some notes and caveats.

  • If your data are formatted. For example, if you have associated "Minnesota" with the number "27", then by default SAS will put "Minnesota" instead of "27". This is easy to solve with a "FORMAT _all_ ;" statement after the BY statement. This will clear all formats.
  • You can obviously include IF and WHERE statements after the SET statement to write out a subset of the data.
  • By default, character values will be left aligned. The magic internet will help you find out how to align character values where you want them.
  • Numbers will not be zero padded after the FORMAT _ALL_ statement, so you may wish to format all numbers with the appropriate Zw.d format you require.

Good luck. Comments are not expected, but welcome.

Posted by eroberts at 5:59 PM

July 11, 2006

1954 Commonwealth mile

7 minutes of your time. One of the greatest mile races in history. It doesn't matter if you know what happened, it still is amazing to see it, and to see film of Landy turning to look the wrong way at the top of the straight, well, just that little better than the still photos of the moment. (Or in sculpture form)

A good trivia question is "who was third?" I didn't hear them say in the video, which appears to have original commentary. The answer can be found at the link.

Posted by eroberts at 12:17 PM

July 5, 2006

Winning and losing

Let me add another twist here: do Democrats want to run Congress--and be blamed for the inevitable gridlock and investigations that would ensue--before the 2008 elections? In terms of the party's long-term interests, wouldn't they be better off letting the GOP stew in increasingly ineffective and unpopular power for the next two years? (from The New Republic)

I quote this New Republic blog entry as representative of a strain of "thinking" you'll come across in too many media outlets. There's some factual basis for this argument. Parties in power, or at least in nominal control of a branch of the U.S. government, tend to lose support. It is unusual when the winning party gets more seats at the next election. It is also true that Americans, like citizens in every other federal democracy, split their votes in the aggregate, delivering control of one chamber to one party, and the other chamber to the other. I say aggregate because there are not a huge number of people who actually vote a split ticket in the same election. It doesn't take a lot of split-ticket voters to deliver different majorities in different races. All that is to say, yes! You can make a plausible case that if the Democrats remain in the minority they'll be more likely to win the Presidency in 2008. And perhaps, perhaps, all around the growing power of the Presidency means that's more important.

But read that quote again: "do Democrats want to run Congress--and be blamed for the inevitable gridlock and investigations that would ensue."

Now, if I'm not mistaken the Republicans have attacked the Democrats for obstructionism and partisanship, and similarly ridiculous procedural behavior, while the Democrats are in the minority. You see, that's politics ... it's like war without the shooting ... Parties will attack each other. You don't avoid that by losing. If anything, a party that persistently loses is easier to attack. I've said before, and I'll say it again, if the Democrats don't win this election they'll be on their way to being like the 1990s Canadian Conservative party or the Australian Labor Party of the 1950s. Being out of power for so long becomes proof to some of the people you're not meant to be in power. Not a happy thing for a political party.

"Inevitable gridlock and investigations"? The investigations are not inevitable, a Democratic Congressional majority decide they don't want to investigate the executive branch, that's a decision they can make after they've won. Nothing inevitable about it.

Gridlock, I will concede, is inevitable if you define gridlock down as being Congress not working for the President as the Republicans seem to have done. But here's the thing. If you believe the polls, Democratic positions on major domestic issues (health, Social Security, the environment) are more popular than Republican ones. Now, see, if the Democrats won control of [a house of] Congress they could propose popular legislation that might be voted down by the other house, or vetoed (heaven forfend!) by this President. And whose fault would that be?? It's a good question, because really Congressional-Executive gridlock is everyone's equal responsibility. But the perception of responsibility, that's different, that's created by the media and by the parties efforts to put their case to the public.

After twelve years out of the majority in Congress, the Democrats need to win to demonstrate they're competitive at the national level, and then to demonstrate that they can wield power in a way the public approves of. That will do more for their 2008 chances than waiting in hope for the Republicans to become more unpopular.

Posted by eroberts at 2:53 PM

July 1, 2006

Summer is over already?!

I know they said the overnight lows were going to drop a little, but a 20% chance of freezing rain? On the 4th of July? At least we can turn the AC off ...

Posted by eroberts at 3:51 PM