Re: Free Software criteria -- about "Software distributions"

[Davi Leal]

MJ Ray wrote:
> Richard Stallman wrote:
> > "Automatically install" is too weak a criterion for what we need to
> > reject.
> >
> > Here's a real life example.  Most systems let the users select
> > packages to install, and most packages are disabled by default.  If a
> > non-free package is disabled by default, then the system does not
> > "automatically install" it.  But its presence in the package selection
> > system is still bad.

I personally agree with the above RMS' statement. Its presence in the package 
selection system is bad because of that attitude will reward schemes that 
take away our freedom, leading to its loss. [1]   Note that only very expert 
users know if such package is free software or not.

 "An ethical approach to computing is critically important in the short
  and long run. As a result of not stressing free software freedoms for
  their own sake, one learns how to lose those freedoms."

[1] 
http://www.digitalcitizen.info/2007/09/27/why-open-source-misses-the-point-of-software-freedom/

> If the non-free package is present in a package selection system, that
> usually means that selecting a free software package may automatically
> install it, in certain circumstances.  So, I think "automatically
> install" is strong enough to cover that real life example.

RMS exposes that just its presence in the package selection system is bad.  He 
does not talk about if such package is automatically installed or not.

I understand the point exposed by the above MJ's paragraph:
  "Distributions should not _automatically_ install non-free packages,
   that is to say, it should warn fist."

However, again, I agree with the RMS point that this is too weak due to:
  The presence of non-free software in the package selection system
  is bad because of that attitude will reward schemes that take away
  our freedom, leading to its loss. [1]

The presence of non-free software in the package selection system is due to 
such organization/distribution has  "_worked to_ provide easy access to such 
non-free software".


> > The criterion we use in the GNU Project is that of "promoting or
> > encouraging the use of non-free software".  This means anything that
> > would tend to lead people towards non-free programs.
>
> I think that is far too broad.  

I personally agree that the 'encouraging' word is not the right word here, due 
to the below MJ's rationale:  for some non-expert people, just bugs in free 
software could 'encourage' the use of non-free software programs which does 
not has such bugs.

> The only free distribution would be a bug-free debian GNU/Hurd,
> if one used that criterion 

I agree at this point with MJ.  I personally think 'encouraging' is not a good 
word, due to the above MJ's rationale.

If some free software program has bugs, people who do not know what free 
software means, could be encouraged to use an equivalent non-free program 
which has not such bugs.

One of the GNU Herds' proposed criteria is:

  "For software distributions,  they must not include any
   non-free software or _work to_ provide easy access to
   non-free software, even if the default is all Free Software."

> and maybe not even that, if mere capability to install non-debian non-free
> software counts as "promoting or encouraging" its use.

Any software distribution allows the download of any kind of software, data, 
etc.

I think the-download-feature is so basic that is must not be considered as 
promotional of nothing.


Just my current personal opinion,
I could be mistaken,

Davi