You know, I’ve grown so accustomed to operating systems with native Unicode support that I’m still getting over the shock that Windows 7 still has I18N holes you could drive a truck through. And this time I’m not talking about the deep-seated belief that all USB keyboards have the same layout.

It would be one thing if you’d just chosen a different encoding, but no, the file system is UTF-16, text files open in UTF-8, and cmd.exe opens in ISO-Latin-1. I can switch cmd into UTF-8 with /u and manually change the output mapping with chcp 65001, but since I can’t select a font containing kanji, that’s of limited utility.

Thanks to the beta Console2 application, I can verify that it’s possible to get, say, a Perl script or a sqlite session to print kanji from cmd.exe, but since Console2’s developers circular-filed a bug pointing out that their app would be completely compatible with Chinese, Japanese, and Korean if the redraw code just counted characters instead of bytes, I don’t expect things to improve any time soon.

(Windows PowerShell, for all its apparent shell power, still runs in a window that can’t seem to display anything but basic missionary-position Western European characters)

I don’t want to layer Cygwin or Ubuntu on top of Windows, for obvious reasons, so I’m forced to resort to Emacs shell-mode to see gorgeous anti-aliased kanji on a bare-bones command-line (well, after I spent 45 minutes deciphering the poor documentation for the arcane font-mapping elisp commands…). Stone knives and bearskins, when any Mac or modern Linux distro supports nine billion languages out of the box.

Amusing note: did you know someone out there is trying to sell a C Shell port for $350 a seat? No, seriously. I’d have choked on my beer if I drank beer.

[and why am I trying to hack Japanese with Perl and SQLite on Windows when I have a perfectly good Mac or six, and a perfectly functional EeePC running Fedora12? Because the Lenovo S12 won’t melt flesh when the CPU gets busy, has a very nice screen and keyboard, and Win7 is in most other respects an excellent desktop operating system. I wasn’t kidding a while back when I said that Perl and Emacs are the difference between a computer and a toy; iPad enthusiasts take note…]

In the Mac/PC wars, I’ve occasionally commented that my primary computer is a Mac because it’s simply more useful to me right out of the box, and it takes less work to add the rest of what I need. Well, a few weeks ago the Lenovo outlet store had a few refurbished (~30% off retail price) S12 netbooks with the nVidia ION chipset that replaces the pathetic Intel shared graphics that the Atom comes with, and while I waited for it to ship, I started assembling things to install.

[Update: in the essential column, add GetGnuWin32, the wrapper for the GnuWin32 packages. Better than CygWin, less conceptually disgusting than Portable Ubuntu]

Bare minimum to make a computer more than a toy (supplied with every Mac):

• Perl (I used Strawberry, not ActivePerl)
• GNU Emacs
• Putty (SSH)
• Sqlite
• Quicktime
• Update: Virtual CloneDrive (mount ISO images)

Equivalent to extremely useful supplied Mac software:

• Safari/Firefox/Chrome
• iTunes
• TrueCrypt
• PasswordSafe
• Sharpkeys (the CapsLock-killer; sadly doesn't install under Win7)
• SecureW2 TTLS (sadly no longer free, despite all earlier versions being GPL'd)

Other stuff that helps make a laptop useful:

• 3G drivers (AT&T USBConnect Mercury)
• Adobe Flash 10.1 (beta, with hw-accelerated video)
• CCCP codec pack (like Perian on the Mac)
• DisplayLink
• GPG4Win
• Mercurial
• Microsoft Office
• Python
• RealVNC
• TeamSpeak
• VLC
• Windows Security Essentials

Fun:

• Steam
• Good Old Games: Fallout, Fallout 2, Might & Magic 6
• World of Warcraft
• Champions Online

I still need to find something that will mount ISO images as file systems, buy a cheap bare drive to use for backups, and bump the RAM from 2GB to 3GB, but I’m set for now. I wish that the ION version of the S12 didn’t replace the ExpressCard slot with an HDMI port, and I’d love to find a Bluetooth mouse that holds up under regular use, but this is a nice little cool-running carry-around machine, with reasonable performance and battery life.

Oh, and I installed the Nanami OS-tan theme that shipped with Japanese pre-orders of Windows 7. :-)

Vista’s generally a nice OS, but the fact that a folder contains JPEGs does not mean that the “details” folder view should include only those data fields that are valid for photographs. Seriously, wtf? “Date taken”, “Tags”, “Rating”, with a tiny little column for file name, and by default no mention of file creation/modification time? And no obvious way to view it as “folder containing files”?

Button 2 broke fast;
replacement eats batteries.
Gosh it's pretty, though.

I just noticed that Word 2007 offers to translate any text in your document, with a combination of on- and offline tools.

Let’s see what it does with some relatively straightforward Japanese prose, namely the first scene of Kyoutarou Nishimura’s murder-mystery story Ame no naka ni shinu…

more...

During my brief vacation, which involved flying to the midwest to see family and catch a really bad cold, I ordered a new laptop. It’s a Mac, of course, because Windows is simply less functional for people like me. I ordered it online from Apple, so that I could get the precise hardware configuration I wanted: 2.93GHz Core 2 Duo, 4GB RAM, 320GB 7200RPM hard drive, and the Japanese keyboard.

Oh, my, the Japanese keyboard. This is going to take some getting used to:

The worst thing about it is that it doesn’t work under Windows. Even a fully-updated Vista Ultimate install insists on treating it as a standard US keyboard layout, and none of the registry hacks or driver overrides you’ll find through Google will help. You know how MacOS X will ask you to press a few keys to help it figure out what kind of special keyboard you’ve attached? Windows doesn’t do that. This is true even for external USB keyboards made for the Japanese market. Apparently the only way to get it to work is to make sure you overrode the keyboard layout during the initial install of Vista/XP. Maybe.

The second-worst thing would be the absence of a “\” key. For historical reasons, the “¥” key replaced it on Japanese keyboards, and you have to type Option-¥ to get “\”.

Third-worst would be the massive slowdown in my typing speed for non-alphanumeric characters, which are in places my fingers don’t know how to find.

Nice things include the 英数 and かな keys adjacent to the spacebar. These handle the input-mode switching, replacing the usual Command-Space toggle. 英数 switches to English/numeric input, かな to kana input. I’m still using the romaji-style kana input, of course, even though this keyboard has a true kana layout printed on it; that’s a project for, well, “never”.

Also, the Control key is where god intended it to be.

Apart from the keyboard, there are no real downsides to the machine. I’m doing a clean migration to get rid of years of cruft, which helps. I think I’ve got all the finicky licensed apps moved over (Aperture and Photoshop adore their new home), and the bulk of the data. I need to reinstall a crapload of Perl libraries and random bits of code, data, and configs, but I’ve got basic functionality. The old MacBook is still under AppleCare for a few more months, so in a week or two I’ll revert it to the factory RAM and hard drive and send it in for some minor repairs I’ve been putting off.

Why is plaintext email converted to HTML on the server? I ask because it’s clear that you’ve never seen what happens when Entourage tries to display the daily log output from a moderately busy mail server. It takes several minutes to render the message, during which time the single-threaded client is completely unresponsive.

I don’t dare kill it, of course, due to the risk of database corruption; I just listen to the fans as they crank up to full speed, and find something else to do.

[side note to the Entourage team: why does resizing the window that’s displaying an HTML-formatted message trigger yet another multi-minute formatting lockup? Did someone hold a gun to your heads and demand that you use an HTML library that was written by four-year-olds?]

I’m once again transcribing written pieces for use in my reading class, and I keep coming across little nuggets of information that I thought I’d gather in one place.

1. Mac Word has two completely separate editing modes, English and Japanese. You can use either language in both modes, but some behaviors differ, and documents originally created in Japanese mode will show their heritage on other, non-Japanese-enabled computers.
2. Switching between them not only requires restarting Word, but locating the "Microsoft Language Register" application inside the Office folder. In Office 2004, you drag the Word icon onto the Register; in 2008, you run it like any other app.
3. In Office 2004, it will always switch you into Kotoeri input mode when you launch Word. Also, installing updates will revert you to English mode.
4. In both, it will change your default settings, including margins, preferred units, and paper size (A4). Once you override these, it doesn't screw them up again.
5. The two most obviously important features you get out of Japanese mode are vertical text (in the Format/Document and Format/Text Direction menus) and furigana (in the Format/Phonetic Guide menu).
6. Do not attempt to type or edit in vertical-text mode; it's like watching paint dry. You can switch back and forth with the convenient Change Text Direction button on the toolbar, or switch to Draft view to edit.
7. Furigana isn't on the standard toolbar. You can open the Extended Formatting toolbar, or bind a key to the FormatPhoneticGuide command (I use Control-Option-P).
8. Don't add furigana until you think you're done with all other editing. The font and size of the furigana are set when they're created (font used for base word, half its size), and words that have been glossed can't be searched for. They're now equations, you see.
9. When you add furigana, Word often supplies the correct kana. If it doesn't know the word, or can't guess the correct reading when you're glossing only the kanji, it will usually default to the first on-reading for each character, but will sometimes just give up. Keep an online dictionary handy, and cut-and-paste between the two windows.
10. Particularly for vertical text, line and page breaks can be very tricky to control. There are two places to tinker: in the Format/Documents menu on the Document Grid panel, and in the Format/Paragraph menu on the "Indents and Spacing" and "Japanese Typography" panels (including the Options window). My usual settings:
    • On: No grid
    • Indentation, Special=First line,14pt (for normal paragraph indents)
    • Spacing, Before=0, After=0, Line spacing=At least,24pt
    • On: Don't add space between paragraphs of the same style
    • Off: Snap to grid when document grid is defined.
    • On: Allow hanging punctuation
    • Off: Allow punctuation at the start of the line to compress (I wish this also suppressed compression of kana at the start of the line...)
    • On: Compress punctuation and Japanese kana (that is, use proportional spacing)
11. For stories, I set the top and bottom margins to 0.75in, and the left and right to 0.5in. That leaves room for headers and footers, which are always printed horizontally.
12. Word's default Japanese font is MS Mincho, which is actually quite nice, but I prefer Apple's Hiragino Mincho at 14pt. I think it's a bit easier for students to make out all of the strokes, and using 14pt sets the default furigana size to 7pt, which is also easier to read.
13. Don't manually select a heavier weight of a kanji font to get bold text; it might work, it might not, and it might appear to work until you print. Just hit the bold button.
14. Always print to PDF, and always check the results out in Preview before really printing. Take particular note of things like vanishing bold text, unexpected compression of character spacing, and punctuation characters that didn't rotate correctly for vertical layout.
15. Specifically: “ ” … : ⁉ ‼ ⁈ ⁇ (and maybe a few others I haven't found yet). For most of these, Word has simply used the Latin font, and you can just force it to use the kanji font. For the quotes, you need to switch from the usual Western style (“”) to the fullwidth straight style with the close at the bottom (〝〟). For the ques-bang combos, though, you can either give up, or insert a very small inline horizontal text box into the column that contains the correct character. Be sure to drag that text box off to your scrapbook for later reuse.
16. Also, Word regularly hoses page numbering when printing; this seems to be tied to the "quick preview" in the print dialog, so wait for it to finish.

More as I’m reminded of them…