Hacker News Reader

Comment by vorg | Hacker News Reader

The name from the article (i.e. 𬎆 ) is included among the 88,000-odd characters in the Unihan repertoire so there shouldn't be any technical problems in using it, only political problems.

View on HN

Just because a character is in Unicode doesn't mean it's supported in practice: it's missing from the font Chrome is using to render your text and is highly unlikely to be selectable from an IME.

I can't see it either with Windows/Firefox :(

I don't have that problem because I've set the Hanazono font (A and B) as first choice for all Unihan characters on my computers (both Linux and Windows). The latest version, released last year , contains fonts for all 88,000 or so characters in Unihan up to extension F. The downside is that the Japanese version of the character is displayed instead of the Chinese where they differ.

Not sure about Mainland China, but usually it's expected that your name can be encoded using the Chinese Commercial code [1] which goes from 0000-9999. Here is an example of an HKID[2]. The 12 digit number is the commercial code.

[1]: https://en.wikipedia.org/wiki/Chinese_telegraph_code

[2]: https://upload.wikimedia.org/wikipedia/en/2/23/HKID_pic-adul...

I paged through 15 sets of characters (at 10 per page) in my Pinyin IME and couldn't find it, and yeah, mine also shows up as a box. Here's the info if anyone's interested: http://www.fileformat.info/info/unicode/char/2c386/index.htm

Ironically, the character is just a box for me, presumably because it isn't supported on my Mac with Safari.

Same (well, box with hex code) on Mac with Firefox.

It's supported, you just don't have a font with a glyph for it.

Well since I'm using the default fonts, wouldn't that effective be the same thing?

No, because supporting a character means knowing all the metadata about it, not merely drawing it on the screen. You have to know whether it's a letter or a number or punctuation so that you can match against it in regular expressions, or do word breaking when you double click on it, etc. You have to know its bidi class so that you can render it in the right direction when putting it on the screen. You have to know the shaping rules so that apply when it's near other characters. None of that comes from your fonts; fonts just have glyphs in them. All the other metadata comes from the Unicode specification.

Many Chinese systems use the GB18030 (in mainland China) or Big5 (in Taiwan and Hong Kong) encodings. Unicode has proven unpopular for complicated, somewhat political reasons.

While in general the Chinese are pretty pragmatic with standards and just go with what seems to be the best out there or is at least well established, in this case it really seems to be somewhere in between. The technical reason was that it is compatible to GB2312 which was the pre-unicode encoding for simplified Chinese, so it made smooth transition of existing systems possible, while at the same time being able to encode the full Unicode spectrum. Some services like baidu actually use a mix of gb18030 and utf8 for different resources which I imagine could be rather strange to manage sometimes.

GB18030 is more or less Unicode, isn't it?

Indeed:

$ echo 张𬎆| iconv -f utf-8 -t gb18030 | iconv -f gb18030 -t utf-8

张𬎆

Reply on news.ycombinator.com