Subject: Re: a tool free software developers need
From: Tom Lord <lord@regexps.com>
Date: Sun, 17 Feb 2002 21:42:05 -0800 (PST)



        > Tom: 
	> [The test platform farm would (will :-) be paid for by
	>  platform vendors.  OSDL is an example of the idea in
	>  action.]


	Brian:

	OSDL is also not a business.  I don't believe you could make a venture
	like this profitable based only on having the hardware or software
	vendors pay for it.

Like a couple of other respondents, I think you're replying to the
platform farm idea out of context: as if I'd proposed the farm as an
FSB in and of itself.  I'm really shocked at how hard it apparently is
to get across what I think is a very simple idea.

In fact, I more or less agree with your assesment of the farm's
viability as a business -- you aren't contradicting me.

Where is there money to be made here?  One place is in engineering and
constructing the platform farm in the first place.  Another place is
in a (new type of) service: proxying for the platform vendors as a
participant in public projects, seeking to deploy 
config/build/test infrastructure improvements that will maximize
return on the continuing costs of the platform farm.

That said, there may even be a little bit of profit available for
maintaining the platform once it's in place (which is why I only "more
or less" agree).  Why?  Well, the cost competitor for the platform
farm is an NPO.  If you can manage the farm at a lower cost than the
overhead of an NPO, there's some potential profit (though surely not a
windfall).  In bidding, if you are an FSB development organization,
you might even be granted a handicap in the maintenance price because
of the extra value of your having expertese in its operation: an NPO
can't later do related custom contracts for just one of the sponsor's;
an FSB can.

       I'd suggest modelling this on the idea that the consumers do
       pay directly for time on the farm - per hour, per platform.

Maybe, as someone else also suggested, there will be customers who
want preferred access and will pay for it.  I'm not sure that even
that is a good idea, though.  Rather, bend over backwards to encourage
third party apps on these platforms.  Give those people you're
thinking of as customers preferred access for free -- just for having
an ongoing relationship with the sponsor's strategic marketing people.
Maybe the simplest implementation of that is as you've suggested: put
a price tag on preferred access, but give to the vendors who pay the
sustaining costs a stack of gift certificates to hand out to third
parties.


     I suspect the most expensive part of the operation will be to pull
     together sysadmins with the wide range of talent you'd need.

I'll bet you're right.  And arranging things so that they don't burn
out on it.  Designing it in the first place is also a task calling for
world-class programming talent.


    I've heard this idea tossed around a couple times in the last 8
    years so if there was a workable business plan there I'd assume it
    would have been explored...

What has changed in only the past couple of years is that GNU/Linux
has become an important part of the business of several big
manufacturers, and an important part of the planning in some adjacent
markets.

Anyway, the platform is only one small part of the transition to a
"Business B" world, which is where this all started.

-t