> From: Dana Powers
> As a general point, what you call "worth" is terribly ambiguous.  Is
> it what a consumer would be willing to pay?  What a producer would be
> willing to sell for?  What a willing producer and consumer would
> actually exchange for?

It doesn't matter. It's a relative measure.

Generally, in your common or garden market for software, the source code
will fetch a higher price.

It is possible to imagine obscure circumstances (a small device with no
resident compiler - or external PC with cross compiler) and an exotic
location (Siberia) in which being in the right place at the right time you
would be able to sell a desperate punter the binary for a lot more than the
source code.

But, as a rule, the software is defined by the source code, and this is more
valuable than just having a binary.

> A producer sells only object code because she finds that more people
> will come back for improvements and support as she is the only one
> with the source code.  This additional revenue is found to be greater
> than the revenue she would have earned had she sold software.  Seems
> very reasonable to me.

It only seems reasonable, IF revenue from selling improvements, support and
binaries is greater than revenue from selling improvements, support and
source code.

1) If the producer has no competitors:

They still have to charge for their labour doing improvements and support
(giving out binaries).

It doesn't matter whether they provide the source code or not because they
have no competitors.

So why wouldn't they sell the source code?
Alternatively, why wouldn't their customers demand the source code upfront
as a condition of the work?

Imagine you are considering approaching a developer who says $500/hr without
source, $600/hr with source. You have a choice. Similarly if they said
$500/hr without, $2,000/hr with.
The source code is always for sale.

Bear in mind the developer wants your money just as much as you want their
labour. If you want the source, and they have no other customers, then
there's little reason they'll price it out of your reach. Similarly they
can't price binary only delivery lower than their labour charge. 

If the producer has no competitors then they have nothing to fear by
enjoying greater revenue from sale of source code. If the customers are
happy with the binaries at a fraction of the price, then that's up to them.
Nothing's stopping them paying for the source code. However, the producer
has a more valuable product (the source code), and has customers comprising
the only market for the source code. Why on earth is the producer going
refrain from selling the source code? 

2) If the producer does have competitors:

Let's suppose that the producer has competitors (the most typical
eventuality).

We assume that some producers will charge for their services, but will
withhold source code (as you fear), and some will include it or price it at
a premium.

Bear in mind that the source code is the work - it does not cost more to
produce source code than the binary derived from it.

The question is, does the producer who withholds their source code have a
market advantage over someone who doesn't?

Can you really undercut the competition by providing binaries only?

"I will give you the binary only, for 50% of my competitors' fee - who also
give you the source code".

Why?

So you're locked in?

The only reason you'd come back is (as you say) to pay for improvements,
which cost the developer more to develop - and they still haven't recouped
the development cost of the work involved in producing the first binary. The
improved binary has still got to be priced cheaper than the competition's
source code.

What's in it for the producer? Why are they giving you the binary at half
the price of their competitors' source code equivalent?

How are they going to recoup this loss?

And if you want the source code, what's wrong with paying the extra? Why
would the developer refuse to sell it to you at any price?

> Hmmm.  The GPL explicitly requires a "no charge license."  
> See clause 2(b).
> http://www.gnu.org/copyleft/gpl.txt

So?

I can walk into a shop buy a copy of Red Hat Linux for $100 and sell it to
you for $50,000. There's nothing in the GPL that says I can't do this, and
nothing that entitles you to take my copy without charge.

Moreover, I can modify my copy and keep my modifications secret, and nothing
in the GPL says I must give my derivative to you, or even sell it to you.
And if I did decide to sell it to you, the GPL doesn't require me to sell it
to you for $0. I can still sell it to you for $5,000,000 if I fancied you'd
pay that much.

Remember: "Free as in free speech, not as in free beer".

> 
> Nonetheless, the GPL quite clearly uses copyright to require source
> code disclosure (See clause 3).  Compare to a BSD-style license, which
> does not.

The GPL doesn't actually require source code disclosure. 

It says that you can either:
1) infringe copyright, or
2) not infringe copyright, or
3) comply with the GPL's terms (which require disclosure of source to
distributed binaries). 

If you don't distribute binaries, there's no obligation to disclose source
code.

In practice there are no repercussions unless you attempt to enforce
copyright, i.e. if you pretend that you are the exclusive copyright holder
of your copyrighted software (that happens to incorporate GPL source code
and violates the terms of the GPL).

> My more general point is that without solid theoretical or empirical
> grounding, we really have no clue what mix of freedom and protection
> would lead to maximum consumer welfare.  Everything is a trade off
> between access and production, and it is just as hard for me to
> believe that no protection is optimal as it is that all protection is.

Consumer welfare?

Protection?

Are you suggesting that 'consumers' should be prevented from modifying
software for their own safety? That this should be properly entrusted to the
developer alone?

How about thinking of everyone as a software engineer?

What is the benefit of denying engineers the liberty to modify each others'
published code?

Why would ALL engineers willingly prefer such a regime?

If any engineer wished to prevent modification of their code they can simply
keep it secret - or if necessary, contract a select few recipients against
such modification.
_______________________________________________
Discuss mailing list
[email protected]
http://freeculture.org/cgi-bin/mailman/listinfo/discuss

Reply via email to