[gfsd] Re: iozone license free or non-free

[Iozone]

Richard,

   While I agree that "Free software" (Where free is the FSF definition)
   is of tremendous societal benefit, and very appropriate for the
   vast majority of applications, I believe that benchmarks are unique
   and due to the nature of being a benchmark there are societal benefits
   from not permitting one to distribute modified versions of
   a benchmark.
   Analogy:
       You apply for a governmental job. You take the standardized
       test and you get your results for ranking. You're done
       and fairness and equality was achieved. You are competing
       on a level playing field.  You are not permitted to modify the
       test, and then redistribute the modified test and then go off 
thinking
       that everything is cool. It's just not fair ...

   Our difference of opinion is a small one.  I agree with you
   that the vast majority of applications could/should be "free", and
   modified revisions being distributed is beneficial to society.
   I disagree on only the specialized case of benchmarks. I believe
   that permitting redistribution of modified benchmarks is not
   beneficial to customers, vendors, or society at large and,
   in fact, is detrimental to society.

   This may also explain why there are so few "Free" benchmarks
   out there, and why folks mistrust the ones that are available,
   and end up going with extremely proprietary benchmarks where
   they can't see the code, don't know what it does, and pay
   a fortune for its use. Iozone bridges this gulf by providing
   a zero cost benchmark where you get the code, and re-distribution
   rights. Everyone is on a level playing field, and can benefit
   from the benchmark and comparable results. The only
   constraints are that folks are not permitted to redistribute
   modified versions of the benchmark, or create modified
   derivatives of the benchmark and go off in the weeds.
   I believe that this is a reasonable constraint for benchmarks
   in general and that benchmarks deserve some allowance for freedom
   to protect the benchmark's integrity, comparability and credibility.

   I don't wish to hinder the FSF direction or goals, which I
   support, and would not be offended if you wish to remove
   Iozone from your distribution lists. If it doesn't fit your
   criteria, then that's cool with me.  I'm just trying to explain
   another perspective of the universe. ( Hopefully without
   creating any rifts or disturbances in the force :-)

   Our difference of opinion is a  philosophical one, with respect
   to how to deal with benchmarks, as a specialized case of
   "Free" software.

   Thus my previous question about a possible new category
   for dealing with the unique nature of benchmarks.

Enjoy,
Don Capps

P.S. It is an honor to be exchanging email with you. I've been
      a fan almost since you were born  :-) Thanks for taking
      the time to chat.



----- Original Message ----- 
From: "Richard Stallman" <address@hidden>
To: "Iozone" <address@hidden>
Cc: <address@hidden>; <address@hidden>; <address@hidden>; 
<address@hidden>; <address@hidden>
Sent: Wednesday, September 13, 2006 2:24 PM
Subject: Re: iozone license free or non-free


>     It may not meet the GPL open source level, but  ALL of
> Iozone is provided at a cost of $0 (zero)  All source code
> is provided for  $0 (zero) and redistribution is at a cost
> of $0 (zero), and I have never even charged anyone for support.
> So... it is "Free"
>
> "Free" in "free software" refers to freedom, not price.
> (See http://www.gnu.org/philosophy/free-sw.html for the definition.)
> So think of "free speech", not "free beer".
>
> Iozone is gratis, but it is not free software.
>
>     but not does not meet the open source requirements.
>
> Our requirements are for free software, not open source.  We have been
> campaigning for the freedom to share and change software since 1983.
> The open source campaign split off from the free software movement in
> 1998, and its definition of open source is derived from our
> preexisting criteria for free software.
>
> For more explanation about the difference, see
> http://www.gnu.org/philosophy/free-software-for-freedom.html.