Unicode is hard; let’s go shopping!

[pne]

The library receipt says that I returned, among other things, Gr¢nne hjerter and De Fortabte sjæles ¢.

Now I wonder how they got ¢ wrong for ø but got æ right.

Comments

[simont]

Code page 437 and code page 850 have ¢ and ø respectively at position 0x9b, but both have æ at position 0x91.

[pne]

I had assumed it was a visual confusion, since the two characters look vaguely similar, but the code-page explanation makes much more sense! Especially given the technology used (or formerly used) in library computers or their receipt printers.

I was also intrigued to read that "the Norwegian and Danish edition [of CP437] replaced cent (¢) with ø, and yen (¥) with Ø." (Which "happen" to be the assignments in CP850.) I hadn't heard of special country-specific editions of CP437; I reminded me of national versions of ASCII (and the concomitant visual mess if you want to program in a braces-and-brackets programming language: if (...) ä aÄiÜ='Ön'; ü, and the like).

[simont]

I hadn't heard of special country-specific editions of CP437

Neither had I, but I suppose it makes sense: once you've had the idea of CP850 in the first place (being partially compatible with CP437 but also containing the full set of 8859-1 top-half characters), it's not much of a further leap to imagine a continuum between the two, of increasing language support on the one hand and increasing backward compatibility with CP437 DOS apps on the other.

It actually never occurred to me that it might be a visual confusion – I went straight for my character set database and grepped it!