Mailing List Archive

Support open source code!

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Please Help with FVWM

>>>>> "Shige" == Shigeru Abe <> writes:

    Shige> Yesterday, I started fooling around with nls and locale. I
    Shige> don't really know what they are for, but they have ja_JP
    Shige> files in there.  I started randomly copying the ja_JP.jis,
    Shige> etc files over the default and noticed some change.  On one
    Shige> particular file, you start getting random kanji instead of
    Shige> english, so that may be the key. I noticed there was no
    Shige> ja_JP.euc in either directories.. do they exist?

This is deep, deep magic.  Apparently the wizards changed their minds
about how to do these things between X11R5 and X11R6.  My XFree86
hierarchy is pure X11R6, and there is no .../nls.  In the JE
distribution, there is no .../locale.  (I always mv the package (ie,
XFree86 in this case) hierarchy to a safe place, then test with a
symbolic link.  If that's OK, I remove the *link*, and unpack the new
distribution fresh.  That makes returning to a working configuration
fairly painless as long as the new distribution doesn't make too many
symlinks automagically.)

I would suggest making sure there is a clean copy of the entire
.../locale directory somewhere easily accessible.  Then look into the
files in .../nls, and find the matching files in .../locale.  The
name schemes are different, and there can be aliases in the .../locale
scheme.  So you have to actually look at the first few lines of each
file to figure out which ones correspond.

There ought to be a way to have multiple files in each locale
directory.  Probably the one named XLC_LOCALE is the one that gets
used, so you could mv the X11R6 one to XLC_LOCALE.X11R6-dist, then the 
JE .../nls/* file to XLC_LOCALE, and things would work out.

The fact that you get random kanji suggests that your locale is being
set improperly.  Try setting and exporting LANG=ja_JP with the
original X11R6 locale files.  (This must be done *before* fvwm is
started in the *shell which starts it*.  While you're playing with
this kind of thing, I strongly suggest changing the end of .xinitrc to

# fvwm
kterm # note, the last program in .xinitrc is run in the foreground

and starting fvwm by hand in that "console".  Remember that exiting
that "console" will also exit X.)  If you copy the X11R5 .../nls/*
files into the .../locale/*/ subdirectories, you will need to fix the
locale.* index files in .../locale/ (I think, as I say this is deep
magic).  Then you can try LANG=ja_JP.ujis.

The locale documentation is pretty weird, and I don't have a recent
version of Slackware installed so I don't know if it got any better.
But you can try looking up locale(1), setlocale(1), setlocale(3), and
locale(7).  The X documentation of locale related functions is
non-existent in the man pages; presumably the Xlib manual is much more 

Apparently in the JE distribution the EUC code is what is called
".ujis"; I have no idea what ".pjis" might be good for.  In the X11R6
distribution, the .../locale/ja/ subdirectory contains the EUC locale.
Apparently the idea is to promote EUC as the most portable.  (The
others are not compatible with mixing languages in the same document.
Apparently Microsoft's shift-JIS is the worst of all.  At least the
Mosaic L10N implementors hate it with passion.)

    Shige> Can it be my computer? "ujis.fvwmrc" was included with
    Shige> every JE I tried and has never worked, but I assume its
    Shige> possible to use otherwise it wouldn't be included in
    Shige> version after version.

I doubt it's your computer, but it could be your environment as I
mentioned above.  You could do a "grep -i 'locale'" on the entire JF

[The rest is more-or-less irrelevant to TLUG; I'm including it because 
 I'm here already and think it's of somewhat general interest.  If you 
 think it's really inappropriate, let me know.]

    Shige> Did Tsukuba ever turn out to become that "Science City"
    Shige> they were talking about in Expo '85 (in Tsukuba)?  I
    Shige> haven't here that much about it, so I was wondering.

Well, as much as any place in Japan can be considered a "Science
City," Tsukuba can |^P.  Don't get me wrong---there is a lot of really 
good science being done here.  But the scientists are more like
ordinary Japanese salarymen than like US scientists, and the culture
is not like that surrounding Stanford, MIT, or even Ohio State.  I
think that one of the reasons that Japan has so few Nobel laureates
compared to the unquestionable talent and hard work of its people is
that the most prominent scientists are all primarily engaged in "grant 
grubbing" and "grovelling before bureaucrats," and really do very
little science.  In fact, that's how you become a prominent scientist
in Japan.  So younger, very talented, scientists also spend huge
amounts of time grovelling (before their seniors as well as before the

This is an extreme portrait, of course, and highly skewed by my
position as a *social* scientist, where the rewards to grovelling and
grant-grubbing are proportionately higher, and the rewards to doing
good, internationally recognized science are proportionately lower,
than in the physical sciences.  But I think that there is definitely a
strong bias in Japan toward social rather than scientific criteria of
success, as compared to the American individual-directed, more apt
to be science-oriented, criteria of success.

This is changing, at least here in Tsukuba---as far as I can tell
Todai gets ever more deeply mired in those attitudes, and Kyodai is
nearly hopeless from the rumors I hear---and the recent (3 years ago)
appointment of Leo Ezaki (Nobel laureate in physics---he's never done
a single bit of research on Japanese soil, please note) as President
of the University of Tsukuba helped a whole lot here on campus.  (He's
bad for me personally---my institute is misplaced in the engineering
college, and Ezaki-sensei wants to rationalize us into a new
behavioral science college, where our base research funding will be
lower.  But that's exactly the right attitude for the University as a
whole, stop the continual reward of people for historical accident,
and increase the rewards for doing current research.)

So in my opinion, it's an uphill struggle, but we are making progress.
Note that I *do* identify myself with this university, do support its
efforts to improve, and do see definite progress.  But I believe that
requires an honest, critical attitude toward certain unfortunate
legacies, and careful consideration of how that goal may conflict with 
Japanese culture as a whole.

                            Stephen J. Turnbull
Institute of Socio-Economic Planning                         Yaseppochi-Gumi
University of Tsukuba            
Tennodai 1-1-1, Tsukuba, 305 JAPAN       

Home | Main Index | Thread Index

Home Page Mailing List Linux and Japan TLUG Members Links