World’s fairs—those grand international exhibitions especially popular a century ago—are best remembered for the installations and architecture they inspired, like the Ferris wheel in Chicago in 1893 and the Eiffel Tower in Paris in 1889.
But if you were actually attending a World’s Fair, your most immediate impression would likely have been something else. “Most people don’t realize this, but music was one of the most ubiquitous features of fairs. Pretty much everywhere you went on the fairgrounds there was some sort of musical sounds,” says Amanda Cannata, a graduate student in musicology at Stanford University whose research on the subject was recently profiled on the university’s news site.
The orchestras, bandstands, and roving musicians weren’t just entertainment: Like the grandiose buildings and national pavilions, the music had a political agenda. At the 1910 Centennial Exhibition in Buenos Aires, an argument about the music program ended with the oligarchic government enforcing a program of elite symphonic pieces before seated audiences, rather than the participatory sing-alongs that the restless population wanted. At the Panama-Pacific International Exposition in San Francisco in 1915, the former US military officials in charge of setting up the exposition’s “Mexican Village” (Mexico itself was pointedly not invited) chose costumes and music that depicted Mexicans as primitive and in need of civilizing, an implicit argument for military intervention in the country’s revolution.
Why computers should guess more
We count on computers to be fast and precise—accurate and exhaustive where our brains take shortcuts. But maybe they’re trying just a little too hard. That’s the suggestion of a team of researchers at Purdue University who have created a new processor, called Quora, which aims to save energy by letting computers relax a little.
The idea, described in the MIT Technology Review, is called “approximate computing.” Everything a computer does—from displaying images to producing sounds to remembering the last time you opened a document—is based on performing mathematical calculations, and those calculations take energy. But not every calculation needs to be performed with exact accuracy to give us the end result we want: A little imprecision is permissible in search, image recognition, and data mining, for example.
The idea has been around for a while, but the problem is that computers have a hard time deciding when it’s OK to approximate and when it’s necessary to be dead on. The Purdue chip could be a step toward solving that. It’s less strict about its calculations in some areas, like handwriting recognition, and as a result it cuts energy use in half overall. Quora isn’t ready for the commercial market, but the MIT article made clear that it’s a matter of when, not if, our computing devices learn what we already know: Sometimes partial effort is good enough.
The EU’s cinnamon wars
Cinnamon seems to be everywhere during the holidays, and the European Union is trying to crack down. On Dec. 20, the Guardian carried an amusing report on a roiling debate about the amount of cinnamon Danish bakers can use to make pastries. The EU recently passed tough cinnamon guidelines in the wake of research showing that a naturally occurring chemical found in the most widely used kind of cinnamon, cassia, can cause liver damage.
This was a crisis in Denmark, whose iconic pastries, called kanelsnegle, depend on a signature cinnamon swirl. In November, Danish food officials ruled that the pastries don’t qualify as “traditional or seasonal foods,” which are permitted to use more cinnamon than normal baked goods. Danish bakers were beside themselves. “We’ve been making bread and cakes with cinnamon for 200 years,” the head of the Danish Bakers’ Association, Hardy Christensen, told the Guardian. “Then suddenly the government says these pastries are not traditional? I have been a baker for 43 years and never come across anything like this—it’s crazy.”