Why people are not switching to Fedora

[Elad Alfassa]

On Mon, May 11, 2015 at 12:16 AM, Ankur Sinha <sanjay.ankur at gmail.com> wrote:
> I've never understood this argument tbh. Instead of us educating
> people on *WHY* the codecs aren't provided by Fedora, I see people
> repeatedly speaking about how not having them in Fedora is a huge deal
> breaker.

It's not the job of an OS to educate people.
Operation systems exists to allow and enable people to do what they
want and need in a safe, secure and efficient manner.

People are, usually, busy. If they install Fedora on a work computer,
they won't want to waste half the workday figuring out why things are
not working, and fixing them.
For some people, software is a hobby, so they have the patience and
time to mess around with it in their free time, to tweak it, to figure
out why it's not working.

Most people are not like that. Most people don't care that much about
software, and never will, and that's fine. They want an OS that "just
works" and let them use their computer the way they need to use it,
without too much fiddling around and fixing issues.

>
> While one reason is patents as someone already mentioned, the other is
> also the philosophy of FOSS - I hope it isn't just about using a free
> of charge system..

https://fedoraproject.org/wiki/Workstation/Workstation_PRD
Fedora Workstation aims to create the best-in-class operation system
for developers. That's what it is about.

Developers probably already know what open source is - we are no
longer the underdog. Open Source solutions are used everywhere. If
you're a developer, chances are you already know what open source is.
If these developers we are speaking about were idealistic about open
source, they'd probably be using Linux already and therefore not
relevant for this discussion... after all, we are asking "why people
aren't switching from Mac or Windows" and not "Why aren't people
switching from Ubuntu".

And the answer to why people are not switching away from Mac or
Windows is simple: Mac OS X and Windows both have one thing in common.
They work. They don't require too much fiddling (especially in the
case of OS X). WiFi works, music works, video works... and even if
Windows or Mac required you to fiddle a bit when you started, you are
already entrenched. Everything works. Installing, learning, and fixing
a new thing would eat away at your precious time, in which you could
be doing something actually productive, or having fun.

>
> I do understand that having multimedia support is important user wise,
> and I'm more than happy to spend time trying to educate people - I'm
> not in favour of any changes that encourage people to install software
> that is not FOSS.

gstreamer1-plugins-bad-freeworld is LGPLv2+. This falls under the
definition of FOSS. Please don't confuse software patents and
closed-source software. They are two very different things.

>
> So, let us help the third party repository become better? Wouldn't
> that be a good way to go to help our users?

Yes, that would be very helpful and it is certainly a way to go.
However, previously when I offered my assistance I got no response
from anyone who could actually point me at where to get started
there... and as such this repository still does not have appdata
afaik.


> Again, it completely comes down to lack of awareness - people that
> have bitten by the broadcom issue (even though broadcom support in the
> kernel has become much much better recently) know better than to buy
> broadcom in the future.
>
So you give people a choice here: "either throw away your brand new
laptop and buy a different one, or don't use Fedora". That's not a
good thing to do. It's safe to assume our target users already have
computers, and that a very large percentage of them have broadcom wifi
chips. If you're a developer with a laptop supplied to you by your
company, you might not have a choice of the type of laptop at all.

Wireless chips are not usually listed in the specs. This means knowing
what you buy requires a lot of research and prior knowledge of which
chips are "good"... Sometimes the information is not available at all.
Sometimes there are multiple editions of the same laptop with
different wifi chips. Sometimes non-broadcom is not an option (if you
buy a Mac, for example).
If someone tries Fedora, and sees that wifi doesn't work, they won't
blame their hardware (they probably don't even know what "broadcom"
is), after all, it worked fine on Windows (or on OS X). They'll blame
Fedora.

Fedora Workstation's main goal is to create a well-integrated OS for
developers. If we keep pretending we can just ignore issues because
they are "hard" or "complicated" then we are not doing a very good
job.

The thing is, we do have things we can do to make the situation
better, instead of saying "no, it's our philosophy" or "this is not a
real issue". We could enable our users to safely get codecs, for
example. Means of doing that have been discussed in this very list in
the past. So yes, purchasing a patent license for all Fedora users is
out of the question... but there are other ways.

For example, including 3rd party repository definitions is still an
option - if I recall correctly the only reason it was veto'd is
philosophical, not legal.

Or we could offer the users to purchase the Fluendo codec pack in an
integrated and secure way. This is possible, and yes, most users
probably won't pay, but at least it's something, and something is
better than nothing.

We could also have an official-ish page with less vague instructions
on how to safely get 3rd party repo to work, with a clear disclaimer
that this is community generated content in such a way that would make
legal happy.

As for wifi, we could implement a pre-install check screen on the live
CD, that will warn you before you install if your hardware has known
issues. If your only network adapter is not supported, you'd want to
know about it *before* you overwrite your main OS. This kind of
utility could even provide a shortlink (so you can write it down) for
instructions on how to use your other OS to get the right drivers or
firmware files and how to install them.

And as a community, we could try and pressure Broadcom (and similar)
to make better (ie. opensource, in the mainline kernel) drivers for
our OS. That sounds much more productive then blaming users for buying
the wrong thing.

These solutions are not perfect. They are also not very easy, but they
are steps we can take. And I'm sure that if you think outside the box
for a bit you can find more solutions.

We have a real potential here, let's not make it go to waste.
-- 
-Elad.