Sunday 11 May 2003

This is a test of Japanese

I’ve implemented Trevor Hill’s Movable Type modifications (explained by Stavros in his comment on my previous post) by:

  • Turning on the PublishCharset UTF-8 and NoHTMLEntities 1 configuration settings in mt.cfg.
  • Ensuring that the character encoding in each of my templates is set to use MTPublishCharset rather than a hard-coded charset i.e. <meta http-equiv="Content-Type" content="text/html; charset=<$MTPublishCharset$>" />
  • Modifying the send_http_header code in lib/MT/ (using Trevor’s code sample from his post in the MT forum).

So now I should be good to go. Here’s the obligatory test post, with Japanese characters (I’m following John’s example and using a proverb). If you have Japanese support enabled in your OS and can see the characters (or not), please leave a comment. (Please note that you don’t have to go through the above rigmarole just to see the characters.)


Language Hat made the excellent suggestion that for visitors who can’t or don’t want to enable CJK support, “it would be good policy to always accompany [Asian text] with transliterations.” (I think I would probably have done that but it’s good to have it as a formal policy.)

So, the transliterated Japanese is San nin yoreba monju no chie, which means “three people together have the wisdom of a Buddha”; or as we would say in English, “two heads are better than one”. A related proverb plays on the fact that the Chinese character for kashimashii (“noisy, clamourous”) is made up of three small versions of the character for “woman”:


Or Onna san nin yoreba kashimashii (“where three women gather, there is a noisy clamor”). As Kittredge Cherry points out in her book Womansword: What Japanese Words Say About Women:

Of all the characters imported from China, [kashimashii] is almost always the first example that springs to mind when linguistic sex discrimination is discussed. Three women add up to a sin worse than noise when the same character is pronounced kan. This spells wickedness or mischief, and it can be stretched into the verb form kansuru, meaning to seduce, assault, or rape. The hidden corollary to the kashimashii character is that a trio of men getting together is nothing remarkable. There is no character composed of three male ideograms. In fact, the male symbol almost never appears as a component of other characters.

Other words reinforce the concept that women can cause a hubbub. In old Japan, the most likely spot for women to gather was beside the well (idobata) where they drew water and washed clothes, so the term “well-side conference” (idobata kaigi) is still used to describe a group of gossiping women. The word for chatterbox (oshaberi), which literally means “honorable talker,” is almost always used to describe—or put down—a woman. Gossip is considered something women do, while there are few similarly derogatory terms for men who babble about trivial topics.

Permalink | Technorati


You two are killing me. I know how to enable CJK (I assume someone will be along with Chinese, anyway) characters, just fine. Then Windows says "show me your Windows CD", and I say "how about this piece of crap 'recovery disk' that's all I got instead?" Time to toss this laptop for something better. How's OS X's support for Japanese?

Posted by Phil Ringnalda on 11 May 2003 (Comment Permalink)

Funny you should ask, Phil, as I was opening the comments to write, "Works a dream in OS X".

Come on, Phil, come over to the light side. Give into the X.

Linguistic sex discrimination. Hope you cover more of this. I was going to say, none of that in English, but of course there is.

Posted by Burningbird on 11 May 2003 (Comment Permalink)

The characters show up fine in Mozilla 1.3 under OSX (10.2.6).

Posted by Hal Rager on 11 May 2003 (Comment Permalink)

They show up just fine in Mozilla 1.4 under FreeBSD as well. The same for Stravos' Korean characters. I didn't even install any special font packages.

Posted by Scott Hanson on 11 May 2003 (Comment Permalink)

And what are you guys (and birds) in the States doing up so late? Go to bed and get some sleep! As for me, I'm up too early on a Sunday morning watching cartoons with my son!

Posted by Scott Hanson on 11 May 2003 (Comment Permalink)

Under Mac OS 8 and 9, Asian language support was near impossible without buying the relatively expensive Japanese or Chinese language kits from Apple. When my family switched to OS X, it was like a dream come true. It has near-flawless language support for dozens of languages, input and output, straight out of the box.

Now if only there were *simple* instructions for Chinese input in RedHat 9. Oh, and Moz 1.3 in RH 9 shows the Japanese characters just fine.

Posted by Micah on 11 May 2003 (Comment Permalink)

Looks good here, Jonathon, as I thought it would. Cool.

Posted by stavrosthewonderchicken on 11 May 2003 (Comment Permalink)

Actually, the language kits were included in OS 9 too, but you had to explicitly install them, whereas they are all there by default on X, along with the nice fonts alluded to before.

Works great in Safari, but not in NetNewsWire's renditon of the RSS feed.

Posted by Kevin Marks on 11 May 2003 (Comment Permalink)

Looks good here. Interestingly, as I have Chinese and Japanese language support installed, the characters that are the same as their Chinese equivalents are rendered in the Chinese font, and then the rest are rendered in a Japanese font. I don't know if it's supposed to do that, but it does make for an interesting way to differentiate between the two, if you were so inclined.

Posted by John on 11 May 2003 (Comment Permalink)

Works on gentoo/linux using Galeon 1.3.3 and Mozilla 1.3, no funky stuff needed. Works out of the box.

Posted by Baldur Bjarnason on 12 May 2003 (Comment Permalink)

Works fine on my OS X, Jonathon, not that I have a clue what it "actually" says.

Posted by Loren on 12 May 2003 (Comment Permalink)

Works fine in Win98SE Moz1.4b, Opera 7.0 and IE6.

Posted by The Dynamic Driveler on 12 May 2003 (Comment Permalink)

Works great in Safari in OS X. Didn't have to do anything to make it happen...apparently the CJK support is on by default.

Posted by Liz on 12 May 2003 (Comment Permalink)

Works fine in IE 6, XP Pro... Interestingly enough, however, it does not work in the RSS feed (, using Sharpreader 0.901.

Posted by Chuck on 12 May 2003 (Comment Permalink)

Fine in Chimera/Camino .7 nightly and also in the netwnewsire 1.02FC1 build. Oh yeah... OS X (10.2.6)

Posted by rick on 12 May 2003 (Comment Permalink)

just to add to the chorus, works fine on Moz 1.3 and Win ME....but then I've had the Windows IME installed for a long time. (and doesn't work in my NewzCrawler RSS reader).

Jonathan, I wonder if it might not be a good idea to point users to how to install the Japanese IME from Microsoft. This page ( has good instructions and download links. (it's from the same site Stavros linked too for the Korean IME)

Posted by Kurt on 12 May 2003 (Comment Permalink)


Posted by ezra on 12 May 2003 (Comment Permalink)

Not sure what you changed, but the RSS feed shows up in NetNewsWire correctly today.

(on OS X, natch)

Posted by Kevin Marks on 12 May 2003 (Comment Permalink)

Under Gentoo Linux it works fine in Knode (which I use to view RSS feeds), but fails in Mozilla. I've set the CJK USE flag and put Mozilla under a recompile, I suspect that will fix it.

Posted by David Dorward on 12 May 2003 (Comment Permalink)

OS X user joining the din of positive response to your Japanese characters.

Posted by AKMA on 12 May 2003 (Comment Permalink)

Thanks everyone for all your responses -- I've written a post summarizing them:


Kevin, the RSS feeds are working now that I fixed the character encoding.

David, that's an incredibly impressive screenshot. How big is your display? (And I'm surprised how nice my text looks -- why doesn't it look that good on my RedHat machine?)

AKMA, you were right all along. The Macintosh (with OS X) rules.

Posted by Jonathon on 13 May 2003 (Comment Permalink)

Works perfectly from Safari at home, but you already knew that. It also works well in Opera 6 on W2K at work. Then again, I've installed some language support options in the past to deal with Cyrillic input, so I may be atypical.

Posted by Phillip Winn on 13 May 2003 (Comment Permalink)

Just a confirmation that the RSS looks fine now in SharpReader too -- well done!

Posted by Chuck on 13 May 2003 (Comment Permalink)

Works in Safari and in NetNewsWire on OSX, and the characters are antialiased, too.

Posted by Lee Phillips on 13 May 2003 (Comment Permalink) - a recompile with different options seems to have fixed it.

The display is 1024x768 on a 15" combined with 1600x1200 on a 19" screen.

Possible reasons that my fonts look nicer:
* My Mozilla is compiled with support for XFT
* I have the rather nice Bitstream fonts installed -
* I have Mozilla set to ignore font suggestions from other sites

Posted by David Dorward on 14 May 2003 (Comment Permalink)

Opera 7.10 on Windows 200 works really well

Posted by Andrew Brown on 14 May 2003 (Comment Permalink)

This Unicode thing rocks. Not just for Japanese character support, but for any that requires (or if it is your personal preference) characters outside the ASCII 128 repertoire.

As Morpheus says, "Here we go":
left single quote: ‘
right single quote: ’
left duoble quote: “
right double quote: ”
em dash: —
en dash: –
three dots: …
euro symbol: €
bullet time: •
pi: π
copyright symbol: ©
registred symbol: ®
trademark symbol: ™

Posted by Kristiaan Thivessen on 18 May 2003 (Comment Permalink)

This discussion is now closed. My thanks to everyone who contributed.

© Copyright 2007 Jonathon Delacour