MingLiU to me and you (and Ming Light to the people most likely to use it)

by Michael S. Kaplan, published on 2009/01/05 10:01 -05:00, original URI: http://blogs.msdn.com/b/michkap/archive/2009/01/05/9274384.aspx

So over in the Suggestion Box, typography comrade Josh asked:

I wonder if you could explore a bit of font naming history and discuss the "Ming LiU" font(s) that is included with Windows. Notice the uppercase 'U' at the end of "LiU"...it is not "Liu", even though many people pronounce it that way.

My semi-educated guess is that it really was intended to be "Ming Li", but was called "Ming LiU" with a capital 'U' as in "Unicode" in order to distinguish it from the original (Big5) version, which may have shipped concurrently with the Unicode version at one time in order to provide HKSCS support. Or maybe it was a developmental/test font that somehow made its way into a release. Or maybe something completely different.

In any case: the current name seems like a mistake on *some* level. If "Liu" was intended, why wasn't it spelled "Liu", with a lowercase 'u'?. If "Li" was intended, what is the deal with that uppercase 'U'? I suspect looking into the Chinese-language name strings in the font may shed more light on this, which I will attempt to do while eagerly awaiting a response from you :-)

Here goes some of my idle speculation/barely professional style research, Josh! :-)

The font in question, MingLiU (and its proportional cousin, PMingLiU)

are the Traditional Chinese staples of Windows.

Of course the ideographs are not proportional in PMingLiU ut the Latins are, as you can see when you look at them side by side:

Okay, getting back to the question in question, after a small descriptive segue....

The Wikipedia article on the Ming typefaces probably explains the characteristics best:

Characteristics
This typeface is characterised, among other things, by the following:

• Thick vertical strokes contrasted with thin horizontal strokes
• Triangular ornaments at the end of single horizontal strokes called uroko (鱗, literally "fish scales") in Japanese, comparable to serifs 
• Overall geometrical regularity
  

...

Possessing variable line weight and characteristic decorations at the end of lines similar to serifs, this type style is comparable to Western serif typefaces, as opposed to the Gothic styles which are comparable to sans-serif.

Often there are number of different ways to write the same Chinese character, they are collectively referred to as variant Chinese characters. Some of those differences are caused by character simplification or word choices, while others are purely orthographic differences such as stroke styling. The styling of the strokes used in the old Song and Ming fonts came from the style used in Kangxi dictionary. After the postwar Kanji reforms in Japan, the most of the Kangxi style characters were considered as Kyujitai (old style), causing newer dictionaries to incorporate two letter styles, or to simply reject the old styles. In modern China, the government uses the new orthographic style, which is incorporated into MingLiU version 5.03 or above. In Japan, dictionary entries offer both new and old fonts. In Korea, popular fonts such as Batang are based on Kangxi style.

Okay, now let's get back to the name. :-)

Although I do have somewhere a Traditional Chinese version of Windows 3.1, it is packed away in a box and even if I did find where it is likely on 5.25 floppies which I don't have a drive for. So I took a look on the network over at MS and found four chunks of file that when spliced together produced a file known as MINGLI.TTF, which the True Type Font Extensions had a few choice things to say about:

now the lack of information may have more to do with the unconventional way I spliced the files together, as is the way it looks in Character Map:

 or the way it shows up in the Fonts folder:

or the Font Picker dialog:

or finally the way the viewer shows it:

We'll leave aside the weirdness where one technology was able to dig out the English font name while the others were not despite the fact that the locale and UI language settings of the machine were English -- or the fact that it picked a PRC encoding rather than a Taiwan one. There is probably a bug or three in there somewhere, though I'm not sure how supported the scenario is! :-)

Note that the name in Chinese for MingLi43 is

細明體

which is also the Chinese name for the MingLiU font in Windows.

With the names in Chinese identical and the older font stuck in an era almost before Unicode and not claiming to be encoded as a Unicode font like pretty much all fonts do these days, the "U" in MingLiU is pretty clearly all about Unicode.

A native Chinese person probably doesn't even realize any of this is an issue, since for them it really isn't -- they have no silly suffixes at the end of the name, something that has been true since the earliest versions of Win9x and NT that sold a Traditional Chinese version (I unpacked those fonts too but they didn't provide a lot of new information).

This naming history figured out above pretty much matches what the Wikipedia article describes:

'Ming Light' (細明體) - Default interface font for Windows 3.0 to Windows XP, derived from DynaLab's DLCMing font family. Originally distributed as raster font in Traditional Chinese version of Windows 3.0, then it was available in TrueType format as 'MingLi43' in Traditional Chinese version of Windows 3.1. Starting from version 2.00, the font was internally sorted in Unicode sequence with Big-5 codepage, and carried the English name 'MingLiU'. In version 2.10, the font file also contained PMingLiU (新細明體). MingLiU was distributed with Traditional Chinese version of Windows 95 to Windows 98, all regional versions of Windows 2000 or later, PMingLiU Update Pack (新細明體更新套件), Traditional Chinese font pack for Internet Explorer 3, Microsoft Global IME 5.02 (Traditional Chinese), Office XP Tool: Traditional Chinese Language Pack. 

So I could have saved the time had i just gone there in the first place, this time at least. But I was on vacation so I'm not too worried. :-)

The last bit I'll talk about is the extra character added for the proportional font. It's Unihan data is interesting:

Calling something that is proportional and that has been around for over 15 years new and fresh and modern is kind of amusing, though perhaps that's just me. After all we still call it New York and New Jersey despite how old they are, so it is no harm if we use such a thing for a font. :-)

This blog brought to you by 新 (U+65b0, a CJK UNIFIED IDEOGRAPH)

# Josh on 5 Jan 2009 11:28 AM:

# Michael S. Kaplan on 5 Jan 2009 5:00 PM:

# John Cowan on 5 Jan 2009 6:28 PM:

go to newer or older post, or back to index or month or day