Subject: Re: the Free Software Movement in Industry
From: Seth Gordon <sethg@ropine.com>
Date: 24 Aug 2001 16:16:43 -0000

	  Still, if they had released each version under the GPL and then
	  depended entirely on consulting and contract-work improvements
	  for their revenue, they would have had a much riskier business,
	  for reasons discussed earlier.

   They would have had a differently risky business, which is not the
   same as a riskier business.

Not necessarily.

I should also point out that changing business models is itself a
risk, since the corporate culture that has adapted itself to one set
of risks may not be as well-adapted to facing the new set.  So if
you're trying to convince a company to radically change its business
model, it's not enough to say "this will reduce your risks"
(especially if you don't quantify how the risks will be reduced) --
you usually have to say "if you don't do things this way, you *will*
lose money".

   The new risk would be offset by other
   factors, such as greater degrees of freedom for encorporating GPL code
   from third parties, 

This is only valuable if there is, in fact, GPL code from third
parties that is worth incorporating.

I think it's no accident that the most successful open-source packages
*started out* as projects that were not critical to the success of the
institution where they were produced.

   the distributed cost benefits of Open Source
   processes,

This is only a benefit if the cost is, in fact, distributed.  Suppose
Software Company X releases its multi-million-LOC telco billing
package under the GPL.  How many other institutions are likely to help
with the package's development?  The most logical thing for customers
to do is to either hire Company X to make improvements (in which case,
the GPL has no added value) or improve the product in-house and *keep*
those improvements in-house (which the GPL permits -- and why should
the customer let its competitors have better billing software?).

The most logical thing for competing software companies to do is
probably *not* to contribute improvements to the product, but to hire
a fresh team of engineers and say: "Download this, test the hell out
of it, and tell us where its performance weaknesses are, so that our
product can compete better."

   and happier customers (whose risks have been reduced).

You haven't quantified the reduction of risk (compared with putting
the source code in escrow, so the customer will get the source if the
vendor goes under), or whether the customer's happiness will translate
into higher profits for the vendor.

-- 
"Rav would never cross a bridge when an idolator was on it; he said, 'Maybe he
will be judged and I will be taken with him.'  Shmuel would only cross a
bridge when an idolator was on it; he said, 'Satan cannot rule two nations [at
once].'  Rabbi Yannai would examine [the bridge] and cross."  --Shabbat 32a
== Seth Gordon == sethg@ropine.com == http://ropine.com/ == std. disclaimer ==