[pp.int.general] Fwd: Re: Why Free Software misses the point

[Fedor Khod'kov]

Andrew Norton <ktetch at gmail.com> writes:

>> With free software, the basic option is to trust a community of people
>> who check each others' work.  This is like science: not perfect, but
>> no better way to find the truth is known.
>
> TRust a community of people who MAY have checked each others work.
> UNLESS they, like everyone else, assumed someone else would do it. And
> those that have checked it, apparently don't look at the code, but
> instead use other pieces of code to check it.

As RMS wrote:
> You also have the option of paying people of your choice to study
> and/or change the program for you.  That way, you choose who to trust.

There is no need to blindly trust free software community (or any other
person, community or organisation), because if you want, you can hire
people to be responsible for the functioning of your software.  The most
obvious example of such people is system administrators, but you can ask
somebody else to provide technical support, or to audit the code -- the
only limiting factor is money you want to spend on such services,
because the number of people who would do such work for you without
being paid for it is obviously limited.

For most people, trust into community's good will and competence plus
personal efforts is enough to ensure what software works properly and
what errors will be fixed, but nobody is prohibited from taking extra
steps.

> If that community says it's safe, and it isn't, what recourse do I have
> for trusting them? None. The only hope is that *someone* spots the bad
> before it harms anything. If it has though, too bad.

It might be possible to sue big corporation for malfunctioning of their
software, but it would definitely require much time, money and effort.
It is unwise to think what such theoretical possibility can provide a
solid basis for trust into software.  Of course, you can hire people to
support or administrate non-free software (and many do so), but their
abilities is always limited, because they don't have full information
and have either to blindly trust software manufacturers or to guess.
With free software, the only limiting factor is their skills, and you
can avoid it by hiring somebody who is more skilled (and whose services
generally more expensive).

> If people would actually start taking legal responsibility for checking
> and certifying free software, that would be a HUGE positive step
> forward. No-one will though, because no-one a) wants to check, and b) be
> put on the hook for other peoples actions.
>
> Thats the crux. No-one wants the responsibility.

People already do so.  System administrators are usually responsible for
software that runs on system they administrate; and in many fields
software certification is a requirement.

With free software, all necessary information is available to anyone,
including people who wants to make checking and maintaining software
(and taking responsibility for its work) their job; there is no need to
make a single person or organisation responsible for all copies of some
program.
-- 
Fedor.