Subject: Re: Take this analogy and fix it (car repair? _not_)
From: "Stephen J. Turnbull" <>
Date: Mon, 7 Jun 1999 22:28:58 +0900 (JST)

>>>>> "rn" == Russell Nelson <> writes:

Stephen J. Turnbull writes:
    rn> Okay, look at car repair.

    >> I don't think this is a correct analogy. [...]

    rn> [...] but [a car] needs to be worked on by a person with special
    rn> skills and training.  The item is one which most people have
    rn> no interest in becoming an expert in, and they rely on it to
    rn> "just work" when they hit the starter switch.

Sounds like audio equipment or TV equipment or a washing machine to
me.  Do any of those really have competitive markets in after service?
(I'll tell you that they don't in Japan, but that's not relevant.)

    >> I would think they make a lot more money on turning out
    >> "Microsoft-certified" tekneeshuns than they do on service.  And
    >> in fact they seem to be quite happy setting up lots of
    >> competition in the repair business.

    rn> Huh?  And do these Microsoft-certified technicians actually
    rn> get access to the source code?  Heck no, all they get to do is
    rn> wax-n-shine the finish and vacuum out the interior.

Source is clearly not relevant to the accuracy of the analogy, since
it's variable in the analogy.  As far as I know, for modern
automobiles you go to somebody who has a fancy analyzer that they
bought from the manufacturer, and technicians trained at some expense
by the manufacturer.  I doubt they have source.  Most of what
auto repair people do is to figure out which subassembly the funny
noise is coming from, then replace the whole thing.  No?

They certainly are not capable of setting themselves up in competition 
with the OEMs; they don't have the heavy equipment.  They're not
generally capable of removing the 4 speed gearing from the tranny and
replacing it with a five speed (unless they buy the whole thing from
the manufacturer).

In other words, from the point of view of what you expect a _software
engineer_ to be able to do _with source_, car repairmen are all wax
'n' shine guys.  No?  MS-trained technicians at least have access to
b*tch files, which are a lot more powerful cutomization tool than
anything available to ordinary car repairmen.

Oh, yeah.  I just noticed that car manufacturers don't need to
restrict knowledge about the schematics of their engines _as long as
the parts are proprietary_ and using non-approved parts voids the
warranty (doesn't it? did the last time I had a car in warranty in the
US, many years ago).

    rn> Plus, you can't underestimate the power of ignorance.  Why
    rn> don't working people agitate against the minimum wage?
    rn> Because they don't know that it hurts them, yet any economist
    rn> will tell you it does.

Well, actually, most working people are mildly to violently for the
minimum wage.  And the fact is, any economist will tell you that it
does _help_ working people, if you define "working" as "currently
employed above the minimum wage."  It's working person wannabees who
are well and truly hosed by the minimum wage, and on average
employables are hurt.

The point is that different people have different interests.
Everybody (except me :-) on this list is a vendor, right?  It's in
your interest that other people open their sources, you've known that
for a long time.  And you have the tools to take advantage of it.  How
unusual, don't you think?

But most users, including managers in large corporate IT shops, don't.
For their mission critical software, they already have a maintenance
contract, with people they (presumably) have vetted and trust.  For
anything else, aren't they basically in the same shoes with Joe Lucky

    rn> Why don't people insist that all software come with source
    rn> code?  Because they don't know it's possible for j. randoms to
    rn> fix problems, yet any programmer will tell you it is.

Even if they did, I don't know that they would insist on it.  _It
might not be worth the effort to raise their voices._  I don't think
my mother would....

Why not?  Because it is _quite_ expensive to hire J. Random to fix a
program, even with source.  He must study the source carefully to find
and understand the bug, as well as the normal operation, even for
relatively well-documented code.  The point is not that this doesn't
beat dealing with a defunct or recalcitrant original vendor hands
down, nor that it wouldn't put constraints on how much they can
charge.  The point is that it has to be a pretty big bug in a pretty
important program before you're willing to invest the resources in
finding an alternative vendor of maintenance services and verify their
competence.  _That's_ why I infer 99% of individual users don't care.

The US government has demanded second sourcing for many years in
military procurement.  It's expensive.  Source code doesn't have the
same upfront expense, it costs nothing to distribute it.  So maybe
people should demand it.  Maybe it would be effective.

But you can bet that firms that think they make their money by holding
code proprietary will charge extra for it, and they'll make you sign
an NDA.  Most of the benefits you're talking about would be captured
by such a system; I don't see real advantage to most users from
demanding _open_ source, especially if they get the impression that
companies demanding big bucks for restricted "source licenses" produce
more featureful products.

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."