🚀 go-pugleaf

[All '>'-ed lines are from
  /gnu.misc.discuss/ utility@quiche.cs.mcgill.ca (Ronald BODKIN) /Mar 26 1990/
unless otherwise attributed.]

Me>And they don't have to [make expensive extensions available to competitors].
Me>Instead, they can pay even bigger $$,$$$,$$$ to
Me>write their own system.

> And they might, given that competitors can use the gift to erode income.

That is why having GNU-based code is more appealing to hardware-producing
companies than to those who only make software.

Me>Yes.  And if the management of _your_ company can see past their noses,
Me>they'd realize that it still makes sense, because other people (and
Me>companies) enhanced software for _your_ company.

> This is "highly illogical".

Notation:
	<upper case letter> :== variable bound to the statement
	<lower case letter> :== constant throughout the proof
	d(X)	:== changes to X 
	c(X)	:== cost of developing X
	t(X)	:== time required to bring X to market
	<<	:== much less than

Symbols:
	s	:== software needed
	g	:== GNU software available
	f	:== a factor relating cost to development time

Axioms:
	 1. c(X) = f * t(X) + o			[cost proportional to time]
	 2. f >= 1
	 3. f is finite
	 4. o = 0				[overhead insignificant]
	 5. c(X + d(X)) = c(X) + c(d(X))
	 6. s = g + d(g)
	 7. t(g) = 0
	 8. t(d(g)) << t(s)

Theorem:
	 9. c(g+d(g)) << c(s)

Proof:
	10. c(X) = f * t(X)			[1,4]
	11. c(s) = f * t(s)			[10]
	12. c(g+d(g)) = c(g) + c(d(g))		[5]
	13. c(g) = 0				[10,3,7]
	14. c(g+d(g)) = c(d(g))			[12,13]
	15. c(g+d(g)) = f * t(d(g))		[14,10]
	16. f * t(d(g)) << f * t(s)		[8,2,3]
	17. c(g+d(g)) << c(s)			[16,15,11]
	    QED

> [...] you can't break a prisoner's dilemna by arbitrarily cooperating.

My point is that there may not be a dilemma in the first place.

> I can tell you, I would
> not be impressed by the management of a company that gave software or
> enhancements to competitors.

I am aware of two companies that use the GNU C compiler as a basis for
their own: NeXT and Data General.  This meant NeXT releasing their
enhancement that made an Objective-C compiler out of the C compiler (a
major enhancement), and DG releasing the first back end for the Motorola
88000 CPU.  Granted, neither company has its financial future assured at
this point, but somehow I doubt it's because of this decision.

> [...] no one in
> their right mind would make an enormous enhancement if the investment
> didn't work.

That's begging the question.

> if GNU is to become a standard, it
> will need to have plenty of flexibility.

Certainly.

Jacob
--
Jacob Gore		Jacob@Gore.Com			boulder!gore!jacob