Welcome to MSDN Blogs Sign in | Join | Help

The history of the path separator in Japanese and Korean Windows

Why is the path separator on Japanese Windows the ¥ character? And why is it the ₩ character on Korean Windows? I've been prodding Michael Kaplan to delve into the history of this quirk, and he finally gave in to my repeated badgering. (Additional discussion on the Korean Won sign, the Japanese Yen sign, and currency symbols in general.)

Published Friday, October 14, 2005 7:00 AM by oldnewthing
Filed under:

Comments

# re: The history of the path separator in Japanese and Korean Windows

Friday, October 14, 2005 10:29 AM by Michael S. Kaplan
I never think of it as badgering. There are just so many things to talk about.

Or maybe I just talk too much. :-)

# re: The history of the path separator in Japanese and Korean Windows

Friday, October 14, 2005 10:37 AM by flyingxu
But why in Chinese Windows, it is '\'? Can you ask Michael Kaplan again?

# re: The history of the path separator in Japanese and Korean Windows

Friday, October 14, 2005 10:51 AM by oldnewthing
Larry Osterman already discussed why the backslash character is the path separator.
http://blogs.msdn.com/larryosterman/archive/2005/06/24/432386.aspx

If you want to ask Michael Kaplan a question, you can do it yourself on his web site.

# re: The history of the path separator in Japanese and Korean Windows

Friday, October 14, 2005 1:24 PM by woops
I've always wondered what made them decide to use \ in the first place.

# re: The history of the path separator in Japanese and Korean Windows

Friday, October 14, 2005 1:37 PM by Matthew Chaboud
This still doesn't really explain why the Yen and Won would end up at 0x5c. It's really what I expected, because it's the answer to most of these questions in computing:

Legacy.

# re: The history of the path separator in Japanese and Korean Windows

Friday, October 14, 2005 4:43 PM by Michael S. Kaplan
Matthew -- these are National standards in Japan and Korea, and they are still used. When we assign code points in Unicode, we usually do not call that "legacy" so we should probably not do that for JIS or KSC. :-)

# re: The history of the path separator in Japanese and Korean Windows

Saturday, October 15, 2005 4:41 AM by Stu
Surely this makes coversion between character sets far harder? How does windows know weather the YEN or WON symbol it encounters is supposed to be a path seperator and converted to 0x5C or an actual currency symbol?
(Supposing we're converting from unicode to Latin-1 or something...)

# re: The history of the path separator in Japanese and Korean Windows

Saturday, October 15, 2005 7:51 AM by mpz
This quirk is more annoying than you may think. In the default MS IME (that comes with Windows XP), if you type in the yen sign, it allows you to henkan it to either U+005C (backslash) or U+FFE5 (fullwidth yen). U+00A5 (the "correct" halfwidth yen) is nowhere to be found.

This is no problem when using non-Unicode applications in a Japanese locale (since for them backslash *is* the halfwidth yen), but when you try to enter the correct yen sign to a Unicode application from the keyboard, you only get backslashes. Of course you can get the correct character from Character Map but it's a pain...

(I suppose there is no way to make the MS IME know whether it's sending characters to a Unicode or a non-Unicode application and decide whether to send a U+00A5 or a U+005C based on that?)

# re: The history of the path separator in Japanese and Korean Windows

Sunday, October 16, 2005 11:34 PM by Norman Diamond
It gets weirder. When is a backslash not a backslash? When you're reading a different part than you're going to read a minute later, in the exact same page that was cited in Michael Kaplan's blog.

Why does Internet Explorer display one as not a backslash and display another as a backslash in the exact same page? Conspiracy minded cynics might guess that Internet Explorer read Mr. Kaplan's wording calling it a backslash and intentionally rendered that one as not a backslash. Ordinary cynics think Internet Explorer isn't smart enough to do that, and give up at this point.

Screenshots:
http://www.geocities.jp/hitotsubishi/backslash_blog.png
http://www.geocities.jp/hitotsubishi/backslash_blog_2.png

# The Old New Thing : On the fuzzy definition of a "Unicode application"

Wednesday, March 15, 2006 10:00 AM by The Old New Thing : On the fuzzy definition of a "Unicode application"
PingBack from http://blogs.msdn.com/oldnewthing/archive/2006/03/15/551897.aspx
New Comments to this post are disabled
 
Page view tracker