An observant reader named Bob Holt noticed a factual error in the following fact:
http://www.gullible.info/archive.php?m=2006-03#post528
• At exactly -40°, the temperature at which the Celsius and Fahrenheit scales are equal, digital thermometers do not work due to a "divide by zero" error.

He writes:
"I work in a laboratory and we have several instruments as well as stand-alone digital thermometers that read in that range. They have no problem whatsoever in reading either temperature."

He's right. Sorry about that. We tracked down the source of the error, and it turns out that the mistake was made by one of the typesetters, and appropriate actions have been taken, let me assure you. (On a completely unrelated note, we now have an opening for a typesetter, so if anyone has experience with a Linotype, please send your resume to the contact listed on the main page.)

The fact should have stated that "most consumer-market digital thermometers" have that problem. This is because most are built with the Texas Instruments TCF212100 chip, which is notorious in the industry for that bug. (TI clearly is embarrassed by it because they mention the bug as only a 5-point type footnote in the errata for that part.) Since that chip has something like 85% penetration of the consumer digital thermometer market, that's why it should have stated "most" and not implied that all digital thermometers suffer from that problem.

I don't know how many industrial-grade digital thermometers use that chip, but its probably not many because it only has a precision of 0.1 degree Celsius.

Again, our apologies and thank you for spotting our error.