[tlug] (off topic) Open source 'leaving Asia behind'

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

Shawn writes:

 > http://www.regdeveloper.co.uk/2007/06/26/open_source_not_asia/
 > 
 > "Open feedback is OK culturally in Western Europe but a big problem in
 > Japan - for example, open criticism can be seen there as a big shame on
 > you"
 > 
 > My experience (not necessarily developing software here) is that
 > Japanese people _are not_ shy about feedback _if_ the issue is clear.

[...]
 > I was stopped dead in my track yesterday when I wrote 2X25=50 and met
 > with derision by a Japanese colleague who insisted and I mean insisted
 > it was wrong.

"hijoushiki".  That's one that didn't get mention as a translation of
"that's no fuckin' joke, man!" but it may well be the most accurate.

Violations of joushiki are a threat to tatemae, hensachi, and the 1955
system, and will not be tolerated.

However, your "not necessarily developing software here" is a critical
caveat that may make your anecdote irrelevant.  As the article points
out, "Software engineering is an art, it's a fundamentally different
mindset to software manufacturing."  In other words, it's completely
hijoushiki.  So criticism of somebody's design is implicitly taken as
a criticism of the whole way they think and work, and is typically
rejected out of hand on that basis.

Also, I don't know anybody who is as good at taking criticism as they
are at offering it, and the Japanese may be a tad more biased in this
way than the common run of humanity, especially because of the whole
mu-ue/me-shita thing.  Cf. the comment about Thailand, or look at the
meat mess in Hokkaido.  "Don't blame us, we were just following
orders.  We need to earn a living too, so it's not our fault."

That said, I think a bigger problem for Japanese OSS is Conway's Law
(a project is doomed to produce architectures that reflect its
organizational structure).  Japanese like tight, "wet" human
relations, so they often get "big ball of mud" organizations, which
(at least in Emacs/Mule and in a lot of "Japanese patches" to standard
text utilities) are reflected in "big ball of mud" implementations.
Japanese prefer to evolve their best practices and pass them on by
word of mouth, rather than codifying them in clear, written form.
Even if there is a clear written form, people are shocked and hurt if
you try to enforce it (cf the high school baseball financial aid
flap).  That's no way to create an API!

I wonder about the willingness to do scutwork that turns a project
into a product, too.  I know that Japan ranks pretty low in Debian's
translation project; there are fewer projects with translation teams
than for almost all languages used in countries with the kind of
penetration of computers that Japan has, and of those that do, the
average number of messages translated is relatively low.  Of course
that may be due to a failure to perceive need, because Japanese are
satisfied with English error messages.

I also think that the badly horked Japanese educational system
deserves a lot of criticism.  Japanese kids today spend a huge
proportion of their time, starting in about 4th grade, simply studying
for entrance exams.  By the time they get to university, the idea that
there is a "right answer" that they should get from a book is deeply
engrained in them.  This wouldn't be so bad, except that it's a binary
standard.  When I try to show students a *better* way, they presume
that what they did was *wrong*, although (cf my banter with Jim J
about C) it's almost always "maladapted" or "inappropriate" rather
than wrong.  But they're not interested in adaptation -- they want to
be taught "the right way", usually as a single tidbit that can be
pasted on the big ball of mud, rather than integrated into a more open
framework.

None of this is universal, of course, but it does mean that the pool
of Japanese who are prepared to go into programming and open source is
even smaller (and weirder! :-) than for us gaijin.