Re: [tlug] [OT]: PHP/MySQL: Japanese input/ouput prgramming advicedocs sought

["Stephen J. Turnbull" <stephen@example.com>]

>>>>> "Jean-Christian" == Jean-Christian Imbeault <jean_christian@example.com> writes:

    Jean-Christian> Quite a few helpful tid-bits in there. Though I am
    Jean-Christian> still hoping that if the data is in EUC-2022 and
    Jean-Christian> only kana MySQL will be able to sort all least
    Jean-Christian> that.

As long as you can count on data being in EUC (you'll probably need a
front-end to deal with browsers that don't tell you/lie about the
encoding coming in, ask Jim Breen about that ;-), and you're willing
to accept JIS order sorts (ie, primary reading for level 1, stroke
count for level 2 and JIS X 0212), you're OK.

For names, get yomi in kana if at all possible, definitely make room
for 8 kana of yomi in the DB (more if they're place names; note you
can save space by stripping the high byte as all hiragana and katakana
are in the same rows, 4 and 5 respectively, this leaves you with
mojibake ascii after or-ing with 0x7F---and they sort correctly as
ASCII!!).  Between yomi for jinmei/chimei and the now familiar JIS
order, you should be in good shape for Japanese-only apps.

Note that the telco uses a (modified, I think) JIS-order sort.  Ie,
names with multiple readings (there are plenty, it turns out) get a
canonical reading, but this is not always JIS order.  Furthermore, all
kanji are ordered by reading, not by stroke count.  This will require
extra effort, and maybe an extra faux-yomi field, but may be
worthwhile when large tables are returned to users.

For more advice than that, bring beer to Tsukuba ;-)

-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
 My nostalgia for Icon makes me forget about any of the bad things.  I don't
have much nostalgia for Perl, so its faults I remember.  Scott Gilbert c.l.py