by Michael S. Kaplan, published on 2012/09/07 07:01 -04:00, original URI: http://blogs.msdn.com/b/michkap/archive/2012/09/07/10347167.aspx
I've talked about vertical text support now and again over the years, for example in blogs like:
There are others, of course. But this is a nice sample that hints at the complexity.
That second-last blog in particular gets into the challenge of trying to imagine what "Vertical Windows" would look like, if such a thing ever existed for Japanese or Mongolian or Phags-pa.
But a comment in that fourth blog by John Cowan showed me that there were some people thinking about the issue, and striving to be more vertical, even if we weren't doing it in Windows:
I looked into vertical Unicode properties a few years back. What needs to happen IMHO is for Unicode to create two new properties: verticality (Top, Bottom, or Neutral) and rotatability (True or False). Verticality specifies whether, when written vertically, the script must appear in a particular direction or not. Generally, CJK characters (hanzi/hangul/kana) are Top, Mongolian ones are also Top, and everything else is Neutral. For example, Latin script often appears vertically on book spines, with English ones always top-to-bottom and German ones normally bottom-to-top: Latin script is readable either way, so it's Neutral.
Contrary to Unicode, Ogham isn't really Bottom, it's also Neutral. An inscription on a stone is read up the side, across the top, and down the bottom if it's long enough, just like a Latin-script inscription on an arch would be. It's just that most inscriptions are short and only have the bottom-to-top portion. In manuscript, it always appears left-to-right. I once asked Michael Everson what would happen if a book title were mixed English/Ogham; would it be written bidirectionally on the spine -- Latin top to bottom, Ogham bottom to top? He didn't think it was a realistic use case.
Rotationality is what you mention above: basically U+4300 ("one") is rotatable, the rest of CJK is non-rotatable, and everything else is rotatable. Given these two properties, all a user has to do is set the overall direction of text progression (the next line can be left of the current line, right of the current line, or below the current line, known in CSS3 as RL, LR, and TB respectively) and all can be done automatically.
I passed all this on to the UTC informally, but nobody expressed an interest.
And another comment by Andreas Goretzky helped me remember that Windows was not as far away from that ideal as we conventionally tend to think:
Thank you for your thoughts... the UI question you brought up is quite interesting.
Before Windows 7 there was absolutely no way to type classical mongolian letters in a simple way. There were fonts for XP and Windows 2000, but you needed special software for getting the right characters, and scissors and glue and a copy machine for rotating the text, and the Menksoft kbd driver for composing the characters. For simplifying the cyrillic typing I made a driver with MSKLC based on the latin transliteration, but for classical script there is absolutely no way. At least not until Windows 7 appeared.
Windows 7 has at least an integrated IME for typing the mongolian script, and it has fonts so horizontally typing and processing works in some apps, and a LIP is announced...... most likely in cyrillic.
Word 2007 can rotate the document, but has no line breaking into the right direction. There are only 3 options, and the fourth is missing. That's about phantasy of product managers. The easiest way for typing that stuff is Excel :-) having grouped cells and rotated document I can "linebreak" by myself.
I was about to give the RedOffice a try, it was founded by the Chinese government to support all of the minorities languages in P.R. China with a word processor.
But the UI is in Chinese by default so I have no chance to find the option. In theory it has the vertical typing feature, it should have vertical menus as well, and it accepts the IME composed text from W7, but without deep chinese language knowledge I can't solve the line breaking and writing direction setup.
If there ever appears a Windows version that has vertical menu and Mongolian script I mark that day in my calendar and celebrate that instead of my birthday.
But most people didn't see it that way.
It's a vision thing.
By which I mean it's a Vision thing.
Because if everyone can see it, then it isn't called Vision.
It's called Sight. :-)
But if you look at Windows 8, you can see us starting to be a bit more aspirational.
Just look at my list of keyboards, based on the Language Profile list:
Do you see those Phags-pa script and Mongolian script entries?
I expect us to find more people trying things out in Internet Explorer and Word and Publisher, seeing what works. And what doesn't.
And envisioning what the future can hold for us....
Feeling inspired yet?
Well I am, and it isn't just a good review that did it for me....
But don't take my word for it. Try things out in Windows 8 (with these products or others) and let me know what you find! :-)
Bruce Rusk on 7 Sep 2012 8:55 AM:
I agree with what John Cowan suggests, except for the exception: "one" is not U+4300 but U+4E00, and it should actually NOT be rotated in vertical text; it remains horizontal whatever the text direction. It's dashes, not the visually similar character for "one," that should be rotated (see the vertical em dash presentation form, U+FE31).
Azarien on 10 Sep 2012 12:22 AM:
Bruce Rusk: That's right, for example, U+30FC (kana long vowel sign) should be rotated, but U+4E00 is always horizontal in shape.