Re: Leaving the guix project

[David Craven]

Hi Ludo,

> I knew I should have stayed on vacation!

Hope you had a nice vacation :)

> The goals haven’t changed since Day 1 though.

I realize that. When I started using/contributing to guix I was not
fully aware of what that meant.

One important thing that has changed for me is that I no longer
believe that not using binary firmware blobs is a better option in
>every< case.

The problem also became real when I actually started using guixsd on
my devices and started considering putting it on other peoples devices
- like my mother's and some other people I help with their IT problems
from time to time.

It is not uncommon for people to be scared into buying a new computer
because "windows XP or Vista is no longer supported" by chrome or
whatever. My goal is to allow people to be able to check their email
and do some webbrowsing on their existing devices.

Using guixsd should decrease my "support" effort not increase it. This
requires that some basic features that users expect work - say
chromium - with all the google SaaS features, and not require users to
buy new devices. The less hardware guixsd supports the higher my risk.
- An old printer doesn't work - the wifi card doesn't work - whatever
else - is a risk for me.

And while I think that the FSDG as it is written for the most part is
a good ideal, it seems that often the "essential freedoms" are not the
only important thing. This also affects other GNU projects. Examples
would be gcc not being modular (creating the need for llvm) and emacs
not supporting lldb. So in practice free software is not software that
respects the essential freedoms, but has a much more narrow definition
which involves unwritten laws no one can read or understand.

Having grown up in an extremely religious environment I have
difficulty accepting religious behavior. While philosophy guides
decision making, to me the difference is when following a strict set
of rules becomes more important than the goal and motivation behind
them.

I've found a new goal of porting rust to riscv and improving llvm
support for riscv, with a whole load of new things to learn. The
barrier to entry to the llvm project seems to be much lower than gcc.
And since I'm learning a lot of "concepts" as I go a modular codebase
is really important...

David