Subject: Re: the walls have ears
From: craig@jcb-sc.com
Date: 28 May 1999 14:34:41 -0000

>But that's not good enough!  What about the freedom *not* to distribute
>source code?  Isn't that also freedom?  Freedom for people is about
>choices, right?

It is my claim that *that* freedom is of no more *overall* value to society
than the difference in value between the freedom of users to have
access to source code of their *programs* over the freedom those users
have to access (freely) source code that was simply *used* in those
programs.

Put another way, from an economic point of view:

  If you convince me that the additional freedoms the GPL gives *users*
  of *programs* have statistically insignificant additional value above
  and beyond the freedoms they'd have to access whatever free source code
  was put *into* those programs...

  ...then you'll have *also* convinced me that the freedoms the non-GPL free
  licenses offer over the GPL, as they pertain to being able to withhold
  the source code for the entire *program* they distribute, *also* are
  statistically insignificant, in which case I won't see the BSD as
  being "more free" in *this* regard than the GPL.

I believe both sets of freedoms *are* statistically quite significant,
since I've been in the industry long enough to know better than to
believe otherwise.  (E.g. being an ex-employee of Prime Computer,
visiting a site using Prime equipment, and not being able to offer
to help them fix a bug in a subsystem I wrote, or understood well,
back when I was an employee.  That's a hypothetical example, but
an easy one to understand.)

But anyone trying to convince me that the "delta value" of end users
to have access to *source* for their *program* is insignificant, but
somehow the value of being able to *deny* that source when *distributing*
that program *is* significant, had better put forward a convincing argument
as to precisely what accounts for that precipitous drop in value.

Otherwise, I'll conclude that they're hypocrites, or perhaps just
not at all concerned for the public good, which is what RMS was,
after all, talking about when he referred to the average freedom of
users of *programs*.

On a different tack, as I mentioned in the taken-private thread, and backed
up by quotes, this particular thread started when somebody claimed nothing
could be "more free" than *BSD*.  RMS responded by simply outlining
*particular* freedoms that, when considered to be valuable, could
put the GPL in a "more-free" light.  It's really sad to see how many
people appear to have willfully misinterpreted what he wrote as saying
"there are no freedoms worth having other than what the GPL gives",
and to have used that misinterpretation as a platform to attack RMS's
integrity and, in at least one private email, my own.  (I've "plonked"
that particular person, something I rarely have to do these days.)

Regardless, the above issue I raise *is* an FSB issue, in the sense
that people who so highly value being able to put free source code
into their proprietary products (or, by indirection, people who
value other's freedoms to do this) are, IMO, *implicitly* valuing
the *inability* of end users to have access to source code for
their *programs* -- to the detriment of those users (when it comes
to those particular programs they have), to the benefit of the
*vendors* who ship non-free *programs* to end users.

FSB's that don't do a good job of looking at things from their
customers' point of view, which includes valuing the freedoms their
*customers* would have if they received their programs under the
GPL, are likely to doom themselves on the Internet.  (That doesn't
mean the FSB must therefore choose the GPL.  It means they must
put a proper *value* on what the customer would receive, additionally,
if the product *was* under the GPL, and be sure to provide that
value, or some reasonable ratio thereof, by other means, if they
choose a non-GPL route.)

Having just watched a CSPAN2 tape of Bezos' "book talk" last night
(he's the creator of amazon.com), it's reasonably clear (to him, anyway)
that the Internet requires a much greater degree of focus on finding,
and serving, the needs of the *customer*.  As he put it, traditionally,
30% of a business would be serving the customer, 70% in marketing
that fact; on the Internet, it's the other way 'round.

So you had better understand what value your customers having your
identical product *under the GPL* would provide them, above and beyond
the proprietary (or even BSD'd, PD'd, AL'd, whatever) license you're
choosing, and be sure to provide them that value some other way,
lest a GPL'ed clone of your product (or a competitor with a better
understanding of that value) come along and snatch them away.

(I'm pretty sure the software "authors" on this already understand,
or at least think they do, the so-called "positive value" to the customer
of *withholding* the source -- "it allows us to put more R&D into the
product, making it better in the future" -- so I won't go into that.
It's taught by most of today's technology culture, including in
universities and so on.  I find even *housewives* understand that
argument.  But until I go through various analogies, like source code <=>
recipe for cake vs. proprietary code <=> actual cake, they usually have
a hard time understanding the value of having the source code to a
given product.  GPL opponents often either don't understand, or
purposefully deprecate, that value, but as businessmen, they take
serious risks when they do so.  Not because of any GPL police; because,
now that GPL'ed software exists, the missing value becomes more visible,
if not to the vendor, to the *customer*.)

Finally, it has been my long-held thesis (well, something between
speculation and rock-solid opinion :) that the Internet is structured
so it can only barely scratch the surface of what incredible power a
GPL-based infrastructure would have.  Attempts to realize this power
via non-GPL-like ("everyone has access to the source for programs")
means, have met with only partial success, IMO for that very reason.

Much of our present infrastructure (including the way we architect
compilers, linkers, object files, etc.) is based on the presumption
that hiding the source code is a *good* thing, and Linux, as wonderful
(and popular) as it is, could be viewed a mere *copy* of that
anti-source-code-based infrastructure.

We're finally getting to the point now where it would be feasible
for a bunch of us GPL types to get together and architect a new
infrastructure (OS and so on) that *assumes* source-code availability
a la GPL, build it, deploy it, and grab enough users to form an
early-adopter user base.

The popularity of Linux having already opened the doors of business
to the GPL model to some extent, the system I have in mind (but
which needs collaboration to build, of course) would have an easier
entree than when I first started suggesting it around 10 years ago.

And, the *value* this system would put into the hands of end users
would be *hugely* greater than the value the present infrastructure
delivers.

So, just as SCO Unix was not the "place to be" as Linux came out of
nowhere, it's wise to not be predicating your business too much on
the existing proprietary infrastructure.  Take from it what you can
while you (think you) can, but don't place long-term bets on your
customers continuing ignorantly or willfully devalue the importance
of their having access to the source code for the programs they're
using.  amazon.com, as well as Linux, has shown how quickly people can
switch from one model to another on the Internet, as a result of
word-of-mouth, which really *can* beat mega-marketing.

        tq vm, (burley)