Miscellania
Liz had the audacity to claim that she had a weird dream last night. It involved attending a wedding-cum-broadway-play in which everything was going wrong. Yawn! I put her in her place ("p0wned", I believe the term is) by going with Bill S. Preston, Esq., Ted "Theodore" Logan, and the guys from my tabletop group to rescue either Bill or Ted's sister from the evil clutches of Demon-Emperor Zod at the satanic "Catholic" school where he was the headmaster. We were dressed as schoolgirls, except for valadil, who we decided couldn't pass for a schoolgirl, so he was dressed as a security guard. Zod tried to trick us with a hallway that forked and a sign directing us left for Emperor Nod and right for Emperor Zod. I don't think we fell for it.

I saw a frog hopping across the shoulder of I-90, looking like he'd just successfully crossed.

The Mass Turnpike's Charlton Service Center is great for three reasons. There's a doorbell in the men's bathroom (pressing it did not appear to do anything), you can get both pretzels and burritos there, and the souvenir penny machine has a sticker explaining why it's legal (because defacing federal currency is okay, as long as it's not for fraudulent purposes, according to 18 U.S.C. §331.

Finally, some quick research into how the Chinese handled telegraphs in the late 19th century. [wikipedia:Chinese telegraph code] indicates that the most common method for transmitting Chinese over telegraph was to use a system where each character was assigned a 4-digit number. I think that the number assigned to a character used a system where each digit indicated something about the structure of the character, but the only system, the four corner method, which I can find documentation for was invented in 1920.

The sender could translate the message to digits, or for an extra fee, the telegraph operator would do it. The message was then sent as a continuous string of digits, which the receiver would decode into groups of 4. There was also a system where each Chinese character was represented by a string of three English characters for greater transmission speed. Sending digits in Morse code is slow; they take five bits each (each decimal digit is exactly five dots/dashes.) Although if they used 10,000 characters — I'm not sure if every value in the coding system they were using corresponds to a character — in an encoding where each character is given equal weight (which isn't really the way you want to go, of course...) it would take between 13 and 14 bits to represent a character, and using the 4-decimal-digit Morse encoding, that's 20 bits per number, which isn't a terrible amount of overhead... I wonder if the fixed bits/digit characteristic of Morse decimal digits helped avoid decoding errors.

I'd've thought that they'd prefer using a romanization to a numeric encoding. The Wade-Giles romanization was invented in 1859 and revised in 1892, so those were around at the time. Cecil Adams's write-up of using romanizations for Chinese makes it sound like it isn't as straightforward a mapping to and from characters as I would've thought, though. It seems that it (well, Pinyin at least, I assume that Wade-Giles is the same) maps to a sound, not a character, so maybe avoiding that ambiguity was worth the trouble of the numeric encoding. I guess that makes sense, you can't really map English letters to anything but sounds... I hadn't thought about it, but it also makes sense that if Chinese has 50k characters, and each character is a single syllable, well, there's going to be a many-to-one mapping between written characters and spoken characters. A system for encoding the appearance of a character (as opposed to the sound of a character) also has the advantage of working across dialects.