Re: Leaving the guix project

[Adonay Felipe Nogueira]

Note: This message is of course not meant to be rude, and not meant to be "the
source of absolute truth", so please take it lightly.

Only for this message, and for the sake of easier understanding, the
term "promote" will be used in this message as synonym for the following:
install, share, sell, teach usage of, recommend.

The problem of promoting non-free software (or functional data, which
covers broader and true object of the movement) is that the promoter
ends up propagating at least one moral dilemma, and worse than that, an
immediate one.

The immediate moral dilemma goes like this: Person A receives (thus
immediate) a software for which he's permitted to use, however, he's
forbidden to share (redistribute non-commercially) copies of the
original work, and after some time, person B asks A if A can share the
software. Average people would usually see two evils, one worse (deny
sharing, and be respect the first agreement), and other lesser (help B,
and break the first agreement). However, as explained in the following
references, free/libre software activists must anticipate this dilemma
by denying receiving the work:

-
  
[[https://www.gnu.org/philosophy/google-engineering-talk.html#freedom-2-moral-dilemma]]
- Previous reference, #freedom-2-spirit-of-good-will .
- [[https://www.gnu.org/philosophy/applying-free-sw-criteria.html]]

Exceptions for these issues apply when non-free software is used as
means to install other free/libre software that is going to be replacing
the non-free one entirely, or when the user is determined to test parts
of free/libre software together with non-free ones so as to help a
complete free/libre replacement be developed (this was how GNU project
itself grew initially), as described at 
[[https://www.gnu.org/philosophy/is-ever-good-use-nonfree-program.html]].

This doesn't mean however, that "free/libre software activists can't use
non-free software", they can, but only for personal use. The risks are
theirs for the taking. Also, one must consider that actions sometimes
convey ideas, so one must make sure to be always moving towards
free/libre computing, when and where such move is permitted, finantially
and technically
([[http://www.fsf.org/resources/hw/endorsement/criteria]]).

As a personal note: my first thought was to simply reply to this
discussion by saying "Do not promote non-free software to people, and
explain to them what is the problem, and you'll be making your support
requests, and our support requests more easy to deal with" (note the
special meaning of "promote"), but I figured that this long comment must
be used instead, otherwise the reply wouldn't be of much use. However,
the original reply is still valid but so: ...

Do not promote non-free software to people, and explain to them what is
the problem, and you'll be making your support requests, and our support
requests more easy to deal with. :)

Also, it's a good idea to tell people to consult you regularly about
hardware, technologies, network services, websites, and software,
specially because they are the nearlest free/libre software activist
they might trust (as long as you remain one, of course). In case of
doubt, you always have general discussion lists such as
libreplanet-discuss. If everything fails, go for the worst case scenario
of "something" being unfriedly to free/libre software movement, or with
"unknown" status of friendliness, at least it's better than making a ruinous 
compromise ([[https://www.gnu.org/philosophy/compromise.html]]).

About LLVM/Clang: Perhaps due to the fact that I'm not a programmer, I can't
understand why on Earth LLVM/Clang even exists, it's like this plethora
of financial/accounting/chat/social-network/whatever free/libre software
that exist and that which none seem to interoperate well with
eachother, or that which none seem to be aiming for the same goals of
the free/libre software movement (although their source code and
licenses are compatible with free/libre software). Even worse, some of
these have the so called "open core" model, and some other harmful
practices like offering different licenses (assumed to be proprietary,
or permissive/non-copyleft) for people who pay for the software, instead
of requiring everyone to pay for the same free/libre software.

Sometimes I receive emails and phone calls from people saying that GCC
doesn't allow "modules", and I tell them: Well, perhaps there is a
mistake in goals here, why would I want a module if I can have
adaptations, alas, it's free/libre software, I can even hire someone to
make adaptations for me. As an important note: if you do some searching
at the wiki of the GCC project, you'll see that there is an effort to
make it modular. Now, still replying to those people who call me from
time to time, I usually end my argumentations by saying the following:
The problem isn't GCC's lack of modularity, quite the contrary, the
problem is on LLVM/Clang's lack of measures to protect software freedom
in the long term for the end/average users, and this has nothing to do
with modularity or no modularity, as can be found by reading the
following references:

- [[https://www.gnu.org/philosophy/applying-free-sw-criteria.html]]
- [[https://gcc.gnu.org/ml/gcc/2014-01/msg00247.html]]

The same argumentations regarding LLVM/Clang and GCC, apply to Emacs not
supporting LLDB.

So as a recommendation to fix the issue, we can do either (based on
[[https://sfconservancy.org/videos/2015-01-15_Bradley-Kuhn_Future-of-Copyleft_LCA-2015.webm]]):

- Do a very important [A]GPL'd optimization to LLVM/Clang project, and
  write/fund researches about such, such that LLVM/Clang project is
  tempted to incorporate those.
- As a personal suggestion: Improve GCC such that it becomes way
  superior than LLVM/Clang. I have seen people saying that GCC and
  LLVM/Clang are at par with each other in everything in regards to
  things similar to speed and optimization.