Chicago sent its river the other way in 1900

I apologised to a self-checkout machine yesterday and it took the news well.

A new habit took 66 days on average in one study

The number that gets repeated is 21 days. The number I can find a study behind is 66. In 2010, researchers at University College London asked 96 volunteers to pick one small behaviour, such as eating fruit with lunch or doing 50 sit-ups after morning coffee, and to repeat it daily for 12 weeks while rating how automatic it felt.

Automaticity rose fastest early on and then levelled off. The average time to reach its plateau was 66 days, and individuals ranged from 18 days to a projected 254, a figure extrapolated because some volunteers were still improving when the 12 weeks ended. Missing a single day made little difference afterwards. Exercise habits took about one and a half times as long to plateau as eating or drinking ones, and every score was the volunteers' own rating of automaticity.

A 2024 systematic review in Healthcare pooled 20 studies and 2,601 participants and found median times of 59 to 66 days, with individual times from 4 to 335 days. Eleven of the 20 studies were rated at high risk of bias, so I hold even that range loosely. A wide spread makes a worse headline than a round number, which may be why 21 keeps getting repeated.

Edison suggested hello for the telephone and Bell recommended ahoy

In August 1877 Thomas Edison wrote to Thomas B. A. David, president of the Central District and Printing Telegraph Company of Pittsburgh. "I don't think we shall need a call bell," he said, "as Hello! can be heard 10 to 20 feet away."

The word was not his. Mental Floss finds it in print in an 1826 issue of the Norwich Courier, where it grabbed attention or expressed surprise and greeted nobody. Alexander Graham Bell, according to Hagley, had recommended "Ahoy, ahoy" for answering the line.

Edison is credited with encouraging "hello" as the answering word, and Mental Floss argues he first imagined callers simply shouting it down the line, with no bell to announce them. The first phone book, published in New Haven in 1878, according to Mental Floss, suggested "a firm and cheery 'hulloa.'" Ahoy lost, which I consider a missed opportunity for every office.

Chicago reversed a river to keep sewage out of its lake

The Chicago River once ran into Lake Michigan, which was also where the city got its drinking water. The fix, finished in January 1900, was the Sanitary and Ship Canal, which sent the river's flow towards the Des Plaines River and the Mississippi watershed instead. The Sanitary District, formed in 1889, spent about eight years on the work, starting in 1892, and built 13 bridges so that roads and railroads could cross it.

Downstream was unimpressed. Missouri sued Illinois and the Sanitary District of Chicago, alleging the canal would "send 1,500 tons of poisonous filth daily into the Mississippi." In 1906 the Supreme Court dismissed the bill, and Justice Oliver Wendell Holmes wrote that "the case proved falls so far below the allegations of the bill."

The defendants answered that many Missouri towns and cities already discharged sewage into those rivers.

A Suzanne Vega song exposed flaws in the MP3 codec

"Everything else sounded quite OK," Karlheinz Brandenburg told NPR, "and Suzanne Vega's voice was destroyed. So we had some work to do." The test, NPR says, took place in 1988, when Brandenburg and colleagues at the Fraunhofer Institute in Germany ran Vega's "Tom's Diner" through their compression system.

Mental Floss notes that audiophiles had been using the track to test speakers since Vega's 1987 album. By his own estimate he listened to the song more than 500 times while developing the codec. I find that oddly moving: the test that mattered most was the worst case they could find, and the average case could take care of itself.

Fraunhofer's own interview with him gives the scale of the challenge. Brandenburg says the brain judges speech particularly critically, and for this track the algorithm reported that it would need four times the available bit rate.

Online ratings form a J, and who reviews may explain why

A tall bar at five stars, a shorter one at one star and almost nothing between: that is the shape Nan Hu, Jennifer Zhang and Paul Pavlou found in Amazon ratings of books, DVDs and videos collected in 2005. Between 72 and 78 per cent of ratings in each category were four stars or higher.

In Communications of the ACM in 2009 they blamed two biases. People who expect to like a product are the ones who buy it, and among buyers those with extreme opinions are more inclined to post, a tendency they call "brag or moan."

Then everyone in an experiment reviewed the same randomly selected music CD. The ratings formed a single peak rather than a J, with only 3 per cent giving it five stars and 7 per cent one star. The authors conclude that most people have moderate tastes, and that the J reflects who chooses to buy and who chooses to report.

The backwards Chicago River is the one I would retell at dinner.