[arch-general] SystemD poll

[Sven-Hendrik Haase]

On 22.08.2012 02:10, Felipe Contreras wrote:
> On Wed, Aug 22, 2012 at 1:48 AM, Patrick Murphy <thegerdur at gmail.com> wrote:
>> Could you give me a brief explanation as to why init scripts are better?
>> I'm newish to Unix style operating systems
> As I said; they are tried-and-true since *decades*, all the problems
> have been ironed out by slow small changes, so if somebody has
> problems they are probably hitting very few people.
>
> Switching to systemd is not a small change, it's a revolutionary
> change, with the potential to break many people's boot (it has broken
> things in Fedora, and openSUSE, and it's happening in Arch Linux as
> well). So, a sensible person would wait until a sensible time to make
> the big switch (which is clearly not now).
>

Arch is not sensible in the conservative sense. Being conservative here 
means waiting for others to make the software more stable. This is not 
really what Arch is about. We regularly move to software that is 
just-about-enough stable to be used. As far as I am concerned, systemd 
is at that point since I was able to convert my laptop to it without any 
problems at all. We also did this with Python 3 and regularly do it with 
all kinds of small packages

We move at this rapid fashion because we are pretty much the snow plow 
of Linux distros. We sometimes break parts in our systems so that the 
ecosystem as a whole can improve more rapidly. If you can't get into 
that, Arch isn't for you. We get to use all the newest software with all 
its shiny new features and the trade-off is that sometimes things break. 
On the other hand, our breakages improve the quality of the software as 
a whole for everybody else. Some upstreams very much appreciate a large 
user base for new releases to iron out any errors quickly.

Arch is also about practicality and pragmatism. If everybody moves to 
some software that is conceptually less simple than currently used 
software, we might move to use that as well. If we do not in such a 
case, it might mean we would have to maintain the old software that no 
ones cares about anymore ourselves which would actually complicate 
matters as suddenly that old software would essentially become and Arch 
project. Think about it like this: In Arch we try to find the best 
meta-simplicity. That is, you don't just need to consider conceptual or 
technical simplicity but also other fairly transparent factors that you 
might only be able to spot if you are more experienced such as 
burden-of-maintenance and community support.