[ih] Internet analyses (Was Re: IPv8...)

[John Day]

I have considered several things I could say in response. But all take too long and become to involved. Basically, I am still debating whether your argument is teleological or argued after the fact so it is self-fulfilling. Not your fault. It is what you saw. There were too many major players trying to protect their existing business and all of them were wrong. The world was changing. The question is did it? I am not so sure.

What I still find perplexing in all of these discussions is why OSI keeps coming up? OSI has been dead for almost 30 years. My conjecture is that it allows this list to avoid looking at the Internet’s equally egregious errors. Heck, the Internet lost the internet layer.  But we don’t want to look at that for fear of what it might uncover, what sacred cows might get skewered.

Take care,
John


> On Apr 27, 2026, at 21:46, John Gilmore <gnu at toad.com> wrote:
> 
> John Day <jeanjour at comcast.net> wrote:
>> Taxpayer money is free.
>> So much for good old American capitalism.
>> State Socialism is much more effective  ;-)
> 
> John, you may think you're being funny, but you don't understand.
> 
> What mattered were the terms on which the software was released.  Not
> the source of the funding.  Freely available software moves markets.
> People readily adopt good products that they can download for free, with
> no "trial versions" and no hidden gotchas.  They think much harder about
> spending money on products.  Most corporate or academic programmers
> don't even have a budget for buying software, they need management
> permission.  Free downloads need no permission.
> 
> DARPA got exactly what they wanted from their contract with Berkeley --
> broad, rapid adoption of TCP/IP in their research community.  That would
> not have happened if they had paid the same amount but had asked
> Berkeley to license the results through a proprietary company.
> 
> At Cygnus, governments were less than 10% of our business.  Capitalist
> companies were paying us to make GCC work better and to release the
> improvements for free.  Having a GCC port for their chip meant more
> sales of their chip.  It meant more satisfied early customers who might
> build it into high volume products.  A dozen chip companies paid us from
> their marketing budgets!
> 
> Other companies paid us for great support for the build tools that their
> internal developers used.  This meant that their next proprietary
> product was more likely to ship on time.  (A 1-week delay in a new
> product from Cisco or Sun could easily cost them $100 million in sales.
> It's easy math when they were selling >$5B of their flagship product a
> year.  It was worth even more if they avoided their competitor shipping
> a next-generation product before they did.  Both companies used our
> tools, and bought our support.)
> 
> Later, Sony paid Cygnus more than a million dollars to build free
> compilers and simulators for the PlayStation.  They made it all back
> dozens of times over when the PlayStation hit the consumer market.  It
> had a bunch of great games, all of which had been developed and debugged
> in our simulator before the hardware was even available.  The competing
> game consoles were released with wimpy games, because the other game
> developers couldn't start a year before the hardware was ready.
> 
> I don't blame folks for not understanding.  As one of the first
> free-software companies, we had to learn all this ourselves.  The freely
> available software stream that we created, was creating value for our
> customers that we hadn't dreamed of when we started out.  Not to mention
> all the value that ordinary users got, from getting our reliable
> compilers for no money and with easy sharing around their organization.
> 
> Free availability of Internet software (and the protocol specs, but
> that's another story) was a significant component of Internet history.
> It reduced the transaction costs of cooperation -- and at its heart, the
> Internet is really a big exercise in voluntary cooperation.
> 
> 	John Gilmore
>