Bizarre Simplified Chinese Font

mr88cet wrote:

The folks who are using the iPad app to learn Simplified, I suspect, either don't know the difference or don't know what to make of the difference.

That makes sense, though I would think that Rosetta Stone would have some folks who already know the characters test out whatever they produce. Hard to believe they would include a Japanese font in the app itself, they have distinctive names like Hiragino, etc., but you never know.

The folks who are using the iPad app to learn Simplified, I suspect, either don't know the difference or don't know what to make of the difference.

They may appreciate at least being asked?

Very interesting discussion.

I read news in a RSS reader (FeeddlerPro) on my iPhone and iPad in 4 languages, English, Simplified and Traditional Chinese, and Japanese. It bugs me so much that there are always some fonts in some languages that look weird.

The problem even appears with Safari.

The problem is worst with bold type fonts.

It seems that iOS is rather dumb in how it selects characters (glyphs) to draw on the screen. There is a priority list of languages (in the Settings app: General : International : Languages). You can re order this list buy choosing a different display language. The chosen language comes to the top.

By default, on standard English phones (e.g, my iphone4 for AT&T), the order is English->Japanese->SimpChinese->TradChinese. When a character is drawn by iOS, it will search for a font, in the above language order, and will use the first font that contains that character. For example, 谢谢各位，the first two characters are available in SimpChinese only, but the last two characters are available in both Japanese and SimpChinese. So iOS ends up picking 谢谢 from SimoChinese, and 各位 from Japanese.

They look close enough, but in a bold type the different is apparent enough to be annoying.

It seems a browser like mobile Safari would use the language hint of a web page to pick the font from the correct language. However, it doesn't. I got the 谢谢各位 example from 6park.com, a Chinese web site, but Safari still uses the Japanese font for 各位.

I think Apple knows this problem. That's why they put Japanese higher than SimpChinese by default. This way Japanese pages look perfect and SimpChinese pages look kind of acceptable. Had this been set in the opposite order, the Japanese pages would look really bad, because the kana characters in the SimpChinese font are truly ugly.

So the solution for iOS 6 is to move your favorite language to the top. But, if you're like me who has four favorite languages, you lose!

I am not sure if Steve Jobs read Japanese, but he definitely did not read Chinese, or else there is no way in **** he would have let this happen.

Tomei Ningen wrote:

It seems a browser like mobile Safari would use the language hint of a web page to pick the font from the correct language. However, it doesn't.

Right, Safari in both iOS and OS X is defective when it comes to this kind of thing.

Since you seem to know the differences pretty well, why not test the other browsers (opera, icab, chrome, atomic, whatever) and let us know if one of them does it properly?

Here is my test page with the 谢谢各位. This page also contains katakana, and there is a link to switch to TradChinese, so you can test it out in your favorite browser.

http://area.6park.com/jp/index.php?app=forum&act=threadview&tid=12411743

I tried Chrome on iPad and it has the same issue as Safari. I suppose it's because Chrome on iOS uses the same web runtime as Safari.

I tried the same page on Nexus 7, and I don't see the 'weird font' problem. It seems Android uses a single 'global font' that supports the entire Unicode range, so there's no dirty trick with font mapping.

Granted, the Android font does not look nearly as good as the fonts used by iOS. It could be because Google was cheap or didn't care about aesthetics, or could be some font design trade offs had to be made to accommodate the 3 汉字 languages.

By the way, I have always felt that Japanese fonts (such as msgothic on winxp) look better than Chinese fonts, especially on low DPI monitors. It seems the Japanese spent more effort in both form and function.

Tomei Ningen wrote:

I tried the same page on Nexus 7, and I don't see the 'weird font' problem. It seems Android uses a single 'global font' that supports the entire Unicode range, so there's no dirty trick with font mapping.

Do you not see any sign of "Japanese" character variants on that device? I can't believe they would not supply a font with those for the Japanese market. Ideally of course it should only show up on Japanese web pages.

Tomei Ningen wrote:

Here is my test page with the 谢谢各位. This page also contains katakana, and there is a link to switch to TradChinese, so you can test it out in your favorite browser.

http://area.6park.com/jp/index.php?app=forum&act=threadview&tid=12411743

In order to test this properly, you need a web page where there is html code telling the browser that the language is chinese, like (html lang="zh"), to see if the browser will use a chinese font even when japanese is higher on the list. I have not been able to find one yet. Do you have any to suggest?

The problem arises because the browser normally has no other way to tell whether some character data is chinese or japanese, it can be either.

PS Here I finally found a page whose code is carefully tagged for language internally. It looks to me like Safari does in fact handle this properly. I don't see any Japanese forms in place of Chinese. Let me know if you see the same results:

http://en.wikipedia.org/wiki/Simplified_Chinese_characters

Tom, you're right. With <span lang="zh" xml:lang="zh"> tags Safari on iOS picks the right font.

Your page is a little hard to verify this, so I wrote a page that shows this more clearly:

http://area.6park.com/jp/index.php?app=forum&act=threadview&tid=12411792

You can see the first 各位 (inside the lang= tag) is displayed in the same font as the 谢谢, but the second one is displayed with a different font.

I'd hope the browser would be more intelligent, and be able to guess the language used -- at least on a non-trivial block of text. Oh well ...

Also, I tested on Nexus 7. It's actually different than what I described above, and is worse than ipad overall.

First, I found that Nexus 7 also does font mapping. It's just not as obvious (as both fonts look crappy). To see the problem, you need to set the system language to be Japanese. Then, with the 谢谢各位 example, if you magnify the page, you can see the font for the 谢谢 has "sharp corners" (Chinese font) and the 各位 has "rounded corners" (Japanese font) -- not sure what the technical terms are for these corners.

Second, on Nexus 7, the lang= tag doesn't seem to have any effect. This can be shown with my test page:

http://area.6park.com/jp/index.php?app=forum&act=threadview&tid=12411792

I tested on Nexus 7, Android version 4.2.1.

Tomei Ningen wrote:

Second, on Nexus 7, the lang= tag doesn't seem to have any effect.

I tried a number of iOS browsers besides Safari -- iCab, Chrome, Atomic, Mercury, Opera. It looked like all obeyed the language tag except for Opera.

The problem is that very few chinese web pages seem to have language tags. As far as I can tell, the sites for the peoples daily and chinese govenment do not, so by default iOS will use Japanese character variants for those simplified characters where these exist.

mr88cet wrote:

If only we could figure out why the "Rosetta Course" iPad app still shows the Japanese-shaped font, even if you've switched your iPad to putting the Chinese-shaped font first in the search list... Well, and then get Rosetta Stone to fix it. :-)

Yes, that does not make sense. Could you post a screenshot of an example, ideally including the character men2 门 ?