Thursday, March 30, 2006

The Sago survivor is a mutant

And I mean that in the best possible way. This man has something different in his genes (emphases mine).
CNN.com - Sole Sago Mine survivor heads for Miracle Road - Mar 30, 2006

I think he's a got a great potential for a complete, possibly complete, recovery,' said neurologist Julian Bailes, who suggested 'genetic individual variability' might help explain McCloy's survival.

Bailes also cited other factors, including that McCloy was about 1,000 feet away from the miners who perished and was 'in better air.'
I wonder if his physicians expected him to walk or talk again. There's something rather unusual about how he managed severe carbon monoxide poisoning. It's unlikely he's unique, so researchers will want to figure out who else responds this way and why. The results could help with managing carbon monoxide poisoning in general.

Yahoo! yellow pages: the problem with being mostly right

Yahoo's yellow pages flopped big time today.

I used it to find a local business. The number was right, and so I sent my wife to the address, using the handy map link.

Wrong.

They moved some time ago. Many miles away. I am in deep doo-doo; ergo so is Yahoo.

From a systems perspective, this is a fundamental problem with a 'mostly correct' solution. Google's algorithmically constructed local search service is even less reliable, but ironically that's not a problem. It's easy to discover that Google's data is stale; so I've never trusted it the way I used to trust Yahoo.

Yahoo's listings have a corporate feel, as though they were updated, validated and maintained. They probably are, but I suspect the paper yellow page listings are still substantially more accurate -- if only because businesses aggressively maintain their paper listings.

Sigh. I hate the paper yellow pages, but maybe I'm stuck with them again. Certainly I can't trust Yahoo's directory service, and the cost of validating what I find may push me back to a solution I thought was dead 10 years ago.

Wednesday, March 29, 2006

Immigration: an interesting debate

Ahh, immigration. It divides Republicans and Democrats alike, so we can get a break from the usual culture wars.

My mother emigrated from England to Canada. I emigrated from Canada to the US. My paternal grandfather emigrated from Ireland to Canada. Speaking of Canada, my birthplace has an interesting take on immigration. They select for wealth and entrepreneurial productivity, a very mercenary approach that favors job growth in Canada and minimizes disruption.

The US has a different approach, and a different problem. Immigrants are selected for a willingness to work in harsh conditions for low wages (often associated with illegal status), for professional rather than entrepreneurial skills and for family bonds to citizens (the latter is where I came in). The consequence is economic benefits to immigrants and their employers, a mildly positive benefit to the nation as a whole, and probably negative effects on some US workers (per a recent Krugman/DeLong set of essays). The calculus is complex; if illegal aliens didn't harvest US crops either robots would do the work or the crops wouldn't be grown here any longer. On the other hand roofers would be paid more -- that work has to be done and can't be outsourced. Nannies would be paid much, much more, but many women and a few men would switch to day care or stop working.

Besides the economic complexities, there are interesting legal and cultural issues. To what extent is the US owned by its citizens -- vs. for example, the foreigners who increasingly own our bonds, our stocks and our land? What special privileges do America's owners demand as a benefit of ownership? Do we owners want to do something to boost wages and employment of less skilled workers, or do we want to boost overall wealth and lessen the impact of the aging boomers?

If it were up to me, I'd take a hard look at what Canada does -- maximize the economic benefits of the immigrant stream. I'd also want to get some real data on the impact on less skilled US workers; I'd probably choose "protection" of some domains. I would also look at a range of measures to favor and increase english language skills; I came from a nation divided by language (Quebec) and I think that's a very bad idea for the US. Lastly, I think a lot of labor intensive agriculture probably doesn't make sense for the US any more.

Tuesday, March 28, 2006

Indict the dogs

Molly Ivins wants to indict military dogs. She also uses the word "plangent". Clearly she is an enemy of America, and now that I've turned her in I assume I'll be forgiven my past sins.

I think Molly is starting to despair. I know I am ...

Monday, March 27, 2006

CEO Corruption: will there be unexpected consequences

Unless someone's made a gross statistical error, it is certain far beyond a mere reasonable doubt that a good number CEO option grants are being backdated to maximize returns:
The Big Picture: CEO Options: Luck -- or something else?:

... A Wall Street Journal analysis suggests the odds of this happening by chance are extraordinarily remote -- around one in 300 billion. The odds of winning the multistate Powerball lottery with a $1 ticket are one in 146 million.

Suspecting such patterns aren't due to chance, the Securities and Exchange Commission is examining whether some option grants carry favorable grant dates for a different reason: They were backdated. The SEC is understood to be looking at about a dozen companies' option grants with this in mind.

The Journal's analysis of grant dates and stock movements suggests the problem may be broader. It identified several companies with wildly improbable option-grant patterns. While this doesn't prove chicanery, it shows something very odd: Year after year, some companies' top executives received options on unusually propitious dates.
What kind of impact does this corruption have on a society? At what point do people start dropping out -- or become receptive to a populist reaction? It's happened in America before.

Prime numbers, Zeta functions, and quantum mechanics

Seed magazine has a very readable article that provides a 200,000 foot view of the relationship between number theory and quantum mechanics: Seed: Prime Numbers Get Hitched. I do wish Du Sautoy had mentioned whether this had any implications for cryptography; I believe current techniques rely in part on the technical difficulty of factoring large numbers. Naively one might think a breakthrough in understanding prime numbers would not be all that great for the stock market.

He mentions by way of background Riemann's role in anticipating general relativity, and also describing the Zeta functions that play a role in both QM and prime number theory. If we do reconcile general relativity and quantum mechanics, it would not be amazing if Riemann should turn out to be a common source. I have a bit of a personal connection here. As a high school student, back before there were photocopiers, I gave a talk on non-euclidean geometry. I don't believe I've subsequently worked my feeble brain as hard as I did preparing for that presentation. I doubt much of the class got anything from it; I could just as well have delivered it in ancient Greek. All the same, Riemann made a lasting impression on me.

At one point in my brash days I foolishly dismissed the uncanny connection between mathematics and physics with some "clever" phrase that I mercifully don't remember. I apologize to my victim. I've long since joined the ranks of those who find the relationship more than a bit unsettling.

Do not use IE until latest bug is fixed

The trick is hackers break into legitimate web sites and then set the trap that uses IE to put bots on home computers. I use Firefox for my own browsing, but at work I need to use IE for internal sites. That's probably OK for now. If you use IE you might consider installing and using Firefox instead for the next week or so. The install is very simple and clean, so it's easy to uninstall or just leave it lying around for weeks like this one.
Security Fix - Brian Krebs on Computer and Internet Security - (washingtonpost.com)

More than 200 Web sites -- many of them belonging to legitimate businesses -- have been hacked and seeded with code that tries to take advantage of a unpatched security hole in Microsoft's Internet Explorer Web browser to install hostile code on Windows computers when users merely visit the sites.

In an update to its Security Response Web log, Microsoft security program manager Stephen Toulouse said the attacks Redmond is seeing against the IE flaw 'are limited in scope for now and are being carried out by malicious Web sites.'

I have to call Microsoft out on both counts, and I think some of what I've uncovered so far about these attacks should make it clear that the situation is serious and getting worse by the hour.
I assume IE on the Mac is safe, but there's not much IE use on Macs any more.

Sunday, March 26, 2006

The myth of the skilled worker shortage

The NYT attacks the hoary nostrum that education and retraining is the answer to layoffs (emphases mine).
Retraining Laid-Off Workers, but for What? - New York Times

... Saying that the country should solve the skills shortage through education and training became part of nearly every politician's stump speech, an innocuous way to address the politics of unemployment without strengthening either the bargaining leverage of workers or the federal government's role in bolstering labor markets.

But training for what? The reality, as the aircraft mechanics discovered, is painfully different from the reigning wisdom. Rather than having a shortage of skills, millions of American workers have more skills than their jobs require. That is particularly true of college-educated people, who make up 30 percent of the population today, up from 10 percent in the 1960's. They often find themselves working in sales or as office administrators, or taking jobs in hotels and restaurants, or becoming carpenters, flight attendants and word processors.

The number of jobs that require a bachelor's degree has indeed been growing, but more slowly than the number of graduates, according to the Labor Department, and that trend is likely to continue through this decade. "The average college graduate is doing very well," said Lawrence F. Katz, a labor economist at Harvard. "But on the margin, college graduates appear to be more vulnerable than in the past."

The Labor Department's Bureau of Labor Statistics offers a rough estimate of the imbalance in the demand for jobs as opposed to the supply. Each month since December 2000, it has surveyed the number of job vacancies across the country and compared it with the number of unemployed job seekers. On average, there were 2.6 job seekers for every job opening over the first 41 months of the survey. That ratio would have been even higher, according to the bureau, if the calculation had included the millions of people who stopped looking for work because they did not believe that they could get decent jobs.

So the demand for jobs is considerably greater than the supply, and the supply is not what the reigning theory says it is. Most of the unfilled jobs pay low wages and require relatively little skill, often less than the jobholder has. From the spring of 2003 to the spring of 2004, for example, more than 55 percent of the hiring was at wages of $13.25 an hour or less: hotel and restaurant workers, health care employees, temporary replacements and the like.

That trend is likely to continue. Seven of the 10 occupations expected to grow the fastest from 2002 through 2012, according to the Labor Department, pay less than $13.25 an hour, on average: retail salesclerks, customer service representatives, food service workers, cashiers, janitors, nurse's aides and hospital orderlies.

The $13.25 threshold is important. More than 45 percent of the nation's workers, whatever their skills, earned less than $13.25 an hour in 2004, or $27,600 a year for a full-time worker. That is roughly the income that a family of four must have in many parts of the country to maintain a standard of living minimally above the poverty level. Surely lack of skill and education does not hold down the wages of nearly half the work force....

... By the spring of 2004, however, out of more than 800 mechanics from United who had gone through her program or were still going through it, only 185 were working again.

Despite their skill, 33 of those 185, or 18 percent, were earning less than $13.25 an hour working in warehouses, on construction jobs, in restaurants or in retailing. Some were "throwing boxes," as the mechanics put it, for FedEx, which paid them only $10 an hour at its shipping center in Indianapolis. They took the work, which entailed loading and unloading air freight packages, for two reasons: FedEx offered them company-paid health insurance, which some of the mechanics desperately wanted, and they saw in the job a gamble worth the hardship, given the glum alternatives.

... Of the 185 mechanics back at work in the spring of 2004, most earned $14 to $20 an hour as heating and air-conditioning repairmen, auto mechanics, computer maintenance workers, freight train conductors (CSX happened to be hiring) or cross-country tractor-trailer drivers, having graduated from a two-week driver-training course offered by Ms. Bucko's people.
Economists have seen this coming for a while. It's likely to get much worse, and this article should be read alongside a recent review of the status of the black American male. Black men are the 'canary in the mine' -- they suffer first.

My prescription? Glad you asked!

1. Universal healthcare in a multi-tiered system. (The universal care is "good enough", not "the best".)
2. Increased taxes on high earners and large asset holders.
3. Reinstate the estate taxes.
4. Universal 401K style savings that can be used both for retirement and for savings. Tax free accumulation on investments, when withdraw you pay taxes at current levels. (Zero if unemployed).
5. Make it easier for people to leave the labor market (see #1).
6. Eliminate any tax features or acccounting rules that in any way encourage outsourcing.
7. Measure what's happening and publish the results.
8. If #1-#7 aren't working, then get radical.

#7 seems obvious, but my recollection is that Bush has eliminated much of this measurement.

The last good toaster?

Toasters are a "missing middle" casualty. China wiped out all the middle and low end products, replacing them with very inexpensive and (as of two years ago) very unreliable toasters. I call this the "missing middle", because afterwords one ends up with the low end (with a great drop in average price) and the luxury/professional market. Alas, I usually buy in the "middle", so this is a bad outcome for me -- even though it's a good thing for most consumers.

In fact, with toasters, it seemed the "high end" had disappeared as well ( commercial toasters too big for our kitchen). I've visited various specialty stores, and the toasters were all made in China there too. They seemed as flimsy and unpromising as the much cheaper models sold in my neighborhood hardware store.

I figured I'd just have to wait until "made in China" came to mean quality products, just as "made in Japan" is today. Or until some retailer rediscovered the value of a "quality brand". I think brands are going to make a big comeback in the next 12 months, so that's not too long to wait.

Today, however, it occurred to me that Germany, with its protected markets and manufacturing inclination might still have reliable toasters. So I changed my toaster search to include "made in Germany". Which led to this Rowenta toaster.

It's a luxury solution of course, I didn't expect to find the "missing middle" in the export market. The electric motor is silly, and the price is steep -- though in the range of the "made in China" items sold in specialty shops. So, at least on the net, there is still a "Mercedes toaster".

Update 3/31: I struck Zeitgeist. Slate has a review of toasters out now. They found an interesting mid-range option. Some of the brands seemed to be going for "quality" too ...

Update 4/14/07: At least some of the components for some Dualit toasters are made in China, though it's "assembled in England". See the comments for more details, including this excellent comment:
Sorry to burst the bubble but the Dualit (at least the Vario 20293 Chrome) may be "assembled in England" but the parts are from China. Check out wholesale site on the internet to see the product country of origin. I bought one, the timer failed after about 3 months, when I opened up the toaster the timer clearly was labeled "made in China". I think that Dualit is assembling them in the U.K. but is using "globally sourced parts" (the new euphemism for Made in China). Buy a cheaper toaster it'll be made in China too but at least it'll say so on the outside instead of on the inside.
Amazon has complaints about early failures in the Dualit. There may simply be no escape from this trap. Maybe we'll all stop eating toast, or go back to the days of holding bread over an open flame.

Update 11/30/07: It's truly hopeless. Bigeejit writes in comments (emphasis mine):
Interesting post! I'm German, and I'm looking for a toaster not "Made in China". It's a nightmare.

Take a closer look at the parts inside of a toaster! Even it's an European or German brand toaster like a Bosch, Siemens, Krups, Rowenta, Braun, Tefal, Moulinex one... most of them look the same, and they are "Made in China".

I believe most of the toasters are manufactured in the same one Chinese plant. Each brand just gets a different plastic cover slipped over.

Well, I'm not willing to pay a high price for a German brand toaster manufactured in China.

@John Gordon: Rowenta is no longer a German brand since 1988. Today the brand is owned by SEB Group (France), and Rowenta toasters are "Made in China", too. Even the Rowenta toaster you mention.
Update 3/17/09: Professional toasters are still made in the US. From comments (Drawde):
.. There is still a toaster MADE IN THE USA. It's made by Star Mfg in MO, actually their toasters are made in TN.

These are heavy duty restaurant quality units. Our 4 slice weighs about 20 lbs, so keep that in mind in terms of storage, lifting etc.

We have the ST04 model, which I think has been replaced with a newer model designation, but I checked with the rep today and she said they are still made in TN.

They can be repaired even at a restaurant repair facility if need be, you can actually buy parts for them as well! Imagine! Of course the bad news, they are not cheap.

I think the 2 slice model is around $350 and the 4 slice $550 online. We actually picked up ours used on ebay for about $150...
Now I know what i want for my birthday ...

Update 6/25/11: This old post still gets comments. Today a vistor suggested toastercentral.com:
... the place to find and buy vintage and collectible kitchen appliances by Sunbeam, Toastmaster, Dominion, Kenmore, Arvin, Westinghouse, General Electric, Manning-Bowman, Universal and other makers from the Golden Age of chrome and bakelite...
The site appeal is primarily aesthetic, but they had a 1950s toastmaster that looked like it would be excellent. Sold of course.

Update 3/3/2012: Six years later, it all makes sense.

Saturday, March 25, 2006

The law of large numbers: identical fake email addresses

Real.com was infamous for abusing users of their "free" audio clients. Spam, nagware, inline obnoxious ads, intrusive reminders, seizing file ownership, quasi-spyware -- the whole nine yards. Bad news.

Alas, the BBC uses RealAudio.

I'd heard RA had reformed out of desperation and litigation, so I downloaded the Mac client. You STILL have to provide an email address, so I provided my usual fake address: nospam@nospam.spa.

It was in use.

So I tried nospam@nospam.nospam.spa.

In use.

Then nospam.nospam@nospam.nospam.spa.

In use.

Finally nospam123.nospam@nospam.nospam.spa

That worked.

I want to know how many clones I have, and how many are a substantial improvement upon me.

Friday, March 24, 2006

The Economist. Adieu.

Twenty years ago I began my subscription to The Economist. I had my first real job, as an intern (aka "first year resident) in family practice. I could afford The Economist.

The first 10 years were superb. Then things began to slip. There were still some fabulous issues though. Their millenium issue was unequalled. Alas, from 2001 onwards it was downhill all the way. The Economist managed to be shocked by Clinton's extramarital affairs; a bit rich for a magazine better known for its libertarian bent. The US coverage grew, the international coverage shrank, and about four years ago they brought in a total moron as "Lexington".

Now they've made the moron their editor.
Crooked Timber � � White smoke at the Economist

... I have to say that my first reaction is to wonder whether it’s too late to cancel the recent renewal of my Economist subscription. I expect the Economist to be vehemently pro-market, but by reading certain kinds of stories with a skeptical eye, and by skipping past certain others, you can find a lot of value in its pages. It has a clear ideological bias, but it isn’t usually actively dishonest. But Micklethwait, together with his scrofulous sidekick Adrian Wooldridge, was responsible for The Right Nation which is one of the lazier and more dishonest books on American politics that I’ve had the misfortune of reading in the last few years, and for the Lexington column which has shown a pretty reliable track record as a purveyor of Republican talking points. There are still a lot of very good people working for the magazine – but I worry that it’s about to undergo a quite substantial deterioration in intellectual quality.

Update: It’s Micklethwait as expected.
CT is too young to know that today's Economist is a mere shadow of its former self. The deterioration is well underway, and now it's a race to the bottom. Happily, over a period of a few years, I've been preparing myself to say good-bye. I'd decided earlier this year not to renew, but I was waiting for the editorial announcement before ordering my alternatives. I will substitute The Atlantic and Scientific American for The Economist. Newsweek has been doing well lately, I'll start reading that online.

Good-bye, The Economist. You were great once, but you have gone the way of all things.

Danger: Addictive temptations from the BBC

I'm becoming a frothing mad lunatic fan of the BBC Radio 4 show In Our Time. I want to buy a T-shirt.

Damnedly, you can't download all the old programs to put them on an iPod. Here's the archive -- full of untouchable goodies for online access only. What am I supposed to on my commute, listen to mere NPR? (I used to like NPR, but now it's fit only to wrap the fish. I've seen the light.)

Maybe it's time for me to get some use out of 'Audio Hijack Pro'. Surely such misuse can't be immoral given the need to elevate my mind.

Let's not even glance at Radio 4's science archives or their history selection. Arggghhh.

Update 3/27/06: I can say that, theoretically speaking, Audio Hijack Pro (AHP), works very well for transferring the archival streams to an iPod. I suggest capturing as 'mono' with a bit rate of 64 kpbs. I use bookmarkable AAC so I can readily return where I left off.

The result is about the same size (20MB) as a podcast. It won't show up in the podcast menu on the iPod (since it's not), but it's easy to create an 'In Our Times' playlist that includes both the shows delivered via podcast and those delivered via AHP. You need the RealAudio client, which is now tolerable.

Read the section in the AHP documentation for tips on configuring capture from RealAudio; you'll do much better than flailing about. Using AHP properly you can set all the tags up in advance, including autonaming the saved file with the title tag using the %name% variable. You can also set the time to 45 minutes and auto-stop on silence. When you're done turn 'record' off and then drag and drop the file onto the "In Our Times" iTunes playlist icon you created.

AHP is slighly flaky. It caused problems on my iBook when the laptop went to sleep with AHP running. Just restart after use.

Theoretically speaking, it's quite easy to build up a a library of a large number of these sessions and catch up on years of IOT one commute at a time. Zoroastrianism is quite good, for example. Actually they all are. I'm awed.

Ask a stupid question ...

A pompous general questions the patriotism of a journalist who writes about national defense. I wouldn't have used his response, but there's something to be said for it:
Early Warning by William M. Arkin - washingtonpost.com:

.... General: 'Mr. Arkin, do you consider yourself a journalist or an American.'

I took a drink of water as my blood boiled.

Me: 'Well General, because I am an American, I cherish the fact that I can call you a f***ing idiot for asking the question.
I'm too genteel to answer that way, and it did get his military sponsors in trouble which seems a bit impolite, but I see the point ...

Graphene, quasiparticles, anyons and quantum braids

Two days ago, I read a popular news report on graphene, a form of carbon with novel properties. In particular it appears to allow one to experiment with quasiparticles, a family which includes not only our familiar bosons and fermions but also, perhaps, anyons. In the article electrons were said to move within the two dimensional graphene surface at relativistic speeds. Here's an article, but it's not what I read.

Yesterday I read an article in Scientific American on topological quantum computing using "braids" to perform error-resistant qubit processing. It's only theoretical, for it to work one needs a good source of anyons. The Sci Am article was written months ago of course. The same issue included a book review pointing in which a physicist saw no challenges to the Standard Model of physics since there are no meaningful places where both quantum and relativistic effects are simultaneously important.

Hmm. Graphene and anyons. - 49 googlits today. Add topological, get 31 hits. Add braid and ... Google stops working. That's funny!

Update: Google finally returned! There were 3 hits when I added braid, including this conference. By the way, most of the physicists mentioned in the SciAm article on using topological methods (braids) to enable error-resistant quantum computing work for Microsofts Project Q. There are surprisingly few googlits on Project Q, apparently it came up during a site visit for one blogger. It's nice to know Microsoft is putting its monopoly rents to interesting uses.

Tuesday, March 21, 2006

Dana Reeve: never smoked, died of lung cancer

The Blog of death covers Dana Reeves, best known, despite her considerable achievements, as the wife of Christopher Reeves.

She was only 44 when she died of lung cancer.

It is worth noting, I think, that she never smoked. As fewer people smoke, what lung cancer there is falls ever more often upon non-smokers. I need to get up to date on what the risk factors are now thought to be. (Second-hand smoke? Radium exposure? ?)