[Mageia-dev] Teamviewer and X86_64 build . . .

Michael Scherer misc at zarb.org
Mon Nov 28 14:40:34 CET 2011

Le lundi 28 novembre 2011 à 12:09 +0100, Kamil Rytarowski a écrit :
> On 28.11.2011 11:32, Guillaume Rousse wrote:
> > Le 28/11/2011 11:28, Kamil Rytarowski a écrit :
> >> There were a few requests for TeamViewer in my local Mageia site - I
> >> suggest to add it to our repositories (probably as tainted and as a
> >> downloading script).

That's non-free, not tainted. 

> > They are better ways to contribute to the distribution than helping 
> > 'poor little users' to install crappy proprietary software. Such as 
> > writing documentation about dependency management, for instance.
> >
> I wouldn't like to follow the OpenBSD way...

Why ?

They managed to be a highly recognized system, seen as one of the best
in his category, and was able to be developed since 1994. 

>  There is software like 
> TrueCrypt, Skype, TeamViewer that many users need. And there is no harm 
> to add it to repositories.

I would like to remind that our focus is doing free software. And it is
not "becoming a dumping ground for proprietary software". 

There is several pragmatic reasons for that and, there is also some long
term harm by shipping more and more proprietary software like : 
- we cannot support them ( no source code, most of the time, no proper
bug report tools, anything ), with the implied consequence of "we cannot
trust them".

- it also make the distinction between application we trust and buggy
stuff that we don't ( skype, flash ) harder to see. We can no longer say
to people "you can trust us, everything in our repository is checked and
supported", since this is not the case. So I personally no longer say to
people to trust us because of that.

- Most if not all proprietary softwares do not permit proper
cooperation, which mean that we cannot plan much around it. So we cannot
place them as proeminent features, unless we want to later risk not
fullfilling our promises ( and I will not talk about how it goes against
cooperation values ). This also place in a uncomfortable position since
we are just treated as 2nd class citizen.

- such software can have a impact on migration to new softwares, and may
requires us to keep old compatibility version, thus adding more QA and
more complexity ( again, I can think of skype, who took a long time to
support pulseaudio, thus forcing more hack to integrate, or the whole
32b/64bits system who was solely needed just for stuff like flash or
wine ). 

- some softwares like nvidia or flash are notoriously buggy, thus making
our distribution look bad when it crash, make people lose time on it
( http://vizzzion.org/?blogentry=819 ), and requires in depth changes to
accommodate them. 

That's just very pragmatic high level points, explaining why we should
avoid them. Please note that I do not even of ethics or anything, just
real life impacts ( impact that people tend to forget to replace them by
philosophical point since that's easier to dismiss as "non important",
but that's just wrong ).

On a more precise point of distributing when the developers do not want
us to do ( like flash, etc ), using a download script is just a ugly
hack as said in the past, since :

- it can break at each change of the website, thus requiring a quick
fix, and sometimes a not so easy one ( what if adobe start to really
protect flash download with a more convoluted approach, like some kind
of JS ? ). So that's fragile and demanding more work. 

- it can be made illegal ( or at least, not authorized by the license
that people never read ). Maybe it is already the case for some
software. Maybe that's not legit to have a contract in electronic form.
While the risk is IMHO rather low, things may changes. I am sure that
Gael Duval didn't think "I could have issues by creating a small
distribution called mandrake". He rightfully thought he would be
protected. Yet, I would rather be cautious to not be too adventurous,
especially for non cooperative software.
As a side note, that would be one reason for me to not propose myself as
president and/or treasurer of the association ( or if I do, . 

- it doesn't work without a working internet connexion. While
flash-plugin and skype are not useful without network access ( so that
point is silly ), some others tools can be, and this would bring more
frustration to people, and more risk of weird errors. I am pretty sure
that people will be puzzled if everything work from their own mirror,
except a few software.

Personally, I aim for better stuff than fragile unsustainable hacks like
this. And we should not start to settle on lower standards, especially
if forced to us by non cooperative company or group. One could be
tolerable. Two, likely. More, that's starting to generalize, and that's
just sending the wrong signal to the whole community, and saying to
company "it is ok to treat us like parias". It is hard to be credible
when saying "we want to provides a solid system" and in the same time,
do the contrary.

I doubt we will be able to bring real long term credibility and a
reputation for quality if we go this way. People can just compare Ubuntu
reputation to Debian one, and see which one was able to survive without
spending million of dollars each year. And also which one serve as a
base for the other.

Michael Scherer

More information about the Mageia-dev mailing list