Subject: cars & code
From: "Stephen J. Turnbull" <>
Date: Tue, 10 Aug 1999 11:00:51 +0900 (JST)

>>>>> "Jean" == Jean Camp <> writes:

    Jean> I wonder if the analogy holds on an industry-wide scale.

I don't see what your analogy is.

    Jean> In the American auto industry General Motors was dominant
    Jean> (gee who would that be in the proprietary SW world...). The

If you're talking about Microsoft, I think that Microsoft's problems
in 2000 are very different from those of General Motors in 1970.  That
part of the analogy, if intended, is very broken.

Of course, extremely large organizations do have inertia.  You could
say that GM's inertia was its downfall, and that MSFT's will be its
downfall.  I agree, but I don't think that's very useful insight now
(it would have been in the 1960s for GM; at that time it wasn't seen
as inertia, it was seen as momentum and considered a good thing[1]).

BTW, did anybody else notice the analysis of stock options in the
Economist this week?  You know, where it was pointed out that if
MSFT's employee stock options were marked to market and the resulting
change in MSFT's liabilities were included as an employee compensation
cost (current FASB accounting rules allow them to be omitted), 1998's
profit of $4.5 billion was actually an $18 billion loss?  (This is a
back of the envelope calculation; the actual expected discounted
liability could be more or less, and would have to take into account
rules about when and how much employees can sell.)

    Jean> management model was highly paid but not very happy
    Jean> labor. Management was seen as stupid and hostile ( Rivethead
    Jean> =? Dilbert).  The model was top-down with captive consumers
    Jean> who had to accept crap, it was just a matter of flavor. Cars
    Jean> were filled with bugs. It was a miracle if the clock in the
    Jean> dashboard worked.

    Jean> The Asian auto industry kicked some American butt (in part
    Jean> because of unfair trade, access to capital, etc) in part
    Jean> because of a networked, distributed knowledge approach to
    Jean> management.

I assume that this is the real analogy to open source that you want to
make.  It certainly has some applicability, but another reason that
Asian industry did relatively well is that knowledge transfer across
corporate boundaries went only one way: in.  Japanese corporations are
famously secretive about everything that matters.[2]  True, the copious
data released about companies is an econometrician's dream---until the
researcher realizes that the definitions of published variables have
little connection to those of interest to either economists or

This "black hole of information[3]" model worked very well during the
catch-up years; the Japanese elite today are famously concerned that
it will not work in the Internet age.

    Jean> Every employee could implement ideas,

True.  The enormous human cost (one of the main uses for the famous
Japanese unpaid overtime was maintenance of machinery -- that was a
matter of pride, and often was unpaid -- and this was one of the main
sources of new ideas; workers in factories twiddling knobs are by
definition not with their families) that paid for this is irrelevant,
I think.  Open software will not create new pressures to split up

However, unlike open software, these ideas were never published, and
often were not formalized, but rather informally disseminated in
quality circles (peer-peer) and mentoring (senior-junior).

    Jean> and the relationship between employee & company as well as
    Jean> customer & company (and therefore emplyee & customer) was
    Jean> very different.

This needs expanding; since the flow of information is different, it's
not clear to me that the customer-company relation is in any way
analogous.  What do you see that I don't?

Also, the Japanese employee-customer relationship is not obviously a
product of Japanese corporate culture; it could easily be based on the
whole culture.  And (like much in Japanese society) extreme employee
-> customer politeness often masks a true attitude which cannot be
expressed in English without four-letter words.

    Jean> Asian cars were cheaper.

But also smaller, less powerful, with fewer options -- at first.

    Jean> There was a hostility or even contempt for Asian autos.

In part a stereotype based on other Asian goods, in part based on the
early attempts to export.  It is not true that Asians just made better
cars (granted, the corporate cultures helped a lot).  Early export
attempts were of no better quality than American goods, and the long
supply lines meant that original sales, and even more so repairs, were
inconvenient and costly to the consumer.  Japanese managers deserve
great praise for recognizing the nature of the problem, realizing that
logistics were the answer to both quality and cost, and finding good
implementations.  Even more so in applying them to areas like design,
which common sense says shouldn't have worked.  But the stereotype was
at least partially justified.  And stereotypes are simply Bayes's Law
in action, so changes in them are going to lag.

This hypothesis (costs of repair drive OEM quality) is strongly
supported by the crap that Japanese consumers routinely put up with in
Japan, crap that even the supine American consumer you describe would
not submit to.

Is the analogous stereotype justified for open source?  Certainly not
on quality of product; so I would guess that the analogy would be to
quality of afterservice (or fears of lack thereof), or perhaps
"featurism".  Here DJ's post is to the point.  (And I wouldn't have
noticed this point without it.)

    Jean> No company would have Asian fleet vehicles.

    Jean> This is one of those ideas that on the surface has an
    Jean> appeal: change in input, change in relationship between
    Jean> producer/manager.

??  I assume you are aware that most early attempts to graft Japanese
practices (quality circles, etc) onto American firms didn't work well, 
despite the fact that many of them originated with American
consultants in the first place.  Also that Japanese attempts to set up 
in the U.S. often ran afoul of cultural factors, eg, equal opportunity 
laws (Japanese companies were famous for a practice analogous to red
lining by banks: they would intentionally locate factories far from
black and Hispanic neighborhoods, and far from public transit).  The
problem is that many separate practices need to be adapted to each

So I think you need to be more specific about the "change in input"
and "change in relationship" that you are talking about.  Or are you
talking about something as broad as the change in viewpoint from
"momentum" to "inertia"?  Not a specific analogy from Asian autos to
open source software?

    Jean> Asian lifetime employment is the opposite
    Jean> extreme from open-source repuation-based returns.

I would argue for a certain analogy here.  Lifetime employment is not
primarily about an insurance contract (although that's the way most
economists seem to see it).  Lifetime employment is about developing
longterm informal human relationships, and I suspect that they can be
viewed as very similar to reputations.

See also, who make a very similar argument.  I
think they are extremely naive about the costs that this kind of
organization imposes, but it does have its attractions.

    Jean> I wonder if any of you would share your thoughts on this
    Jean> comparison? Is it useful and does it offer insight? Would it
    Jean> be a useful or confusing teaching tool?

Well, I wouldn't use it on my (mostly Japanese) MBAs.  They would be
very confused about the idea of spilling information, the company's
lifeblood, onto the street as "open source."  And certainly would not
recognize any analogy to their own previous work experience[4] ;-)

Speaking for myself, I would say that the main lesson I would draw
from the analogy is simply that a given big company can't dominate its 
industry, let alone the whole economy, forever.  But I think you want
more than that.

[1]  Although not by everybody; cf whoever invented Theory Y companies 
(Peter Drucker?) and Robert Townsend, _Up the Organization_.  Not to
mention classics like _The Organization Man_ and Galbraith's stuff.
But this was not conventional wisdom, certainly not in GM's board room.

[2]  A fair amount of information sharing (among Japanese firms only)
does take place under regulatory duress, but the quid pro quo
historically was assistance in export markets and encouragment of
domestic cartels, including protection from imports as necessary.

[3]  The former Minister of Education I am paraphrasing actually wrote 
"black hole of culture," and was referring to Japanese society itself.

[4]  But then, "analogies to work experience" is not a subject taught
in entrance exam cram school, nor in "company spirit" training.  They
miss a lot of the fun stuff :-(

University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
What are those two straight lines for?  "Free software rules."