1 |
On Sun, Feb 23, 2014 at 7:35 AM, Mick <michaelkintzios@×××××.com> wrote: |
2 |
[ snip ] |
3 |
> I am not sure if people object to the Lennart-way of messing up Linux, under |
4 |
> the blessings of RHL, or if they just don't like the immediate outcome. |
5 |
|
6 |
Actually, most people that actually *try* using systemd and reads how |
7 |
it works have no problems with it, and of those there are many (like |
8 |
me) who actually quite like it. |
9 |
|
10 |
> Essentially, in his arrogance Lennart only needs to code things the way *he* |
11 |
> sees as useful or expedient to him and his pay masters. In doing so he throws |
12 |
> the *nix way of developing software out of the window and creates a convenient |
13 |
> for him monolith. Wherever he can't be bothered to do a neat and versatile |
14 |
> job he makes his own arguably option-limiting decisions and thus we have |
15 |
> arrived to today's flavour of systemd-udev-pulseaudio-gnome and whatever else |
16 |
> he will try to weld in tomorrow. He found like minds in Sievers et al and |
17 |
> money from RHL helped them get there. |
18 |
|
19 |
And he also found like minds in some of the kernel developers, and |
20 |
some people from OpenSUSE, and Arch, and Debian, and Gentoo, and even |
21 |
Ubuntu, and old Linux gurus like Keith Packard and Neil Brown[1]. |
22 |
|
23 |
> It ain't pretty and architecturally does not follow the *nix design |
24 |
> principles, but as Canek says, those who can code better should step up to the |
25 |
> plate and redesign systemd as it should have been done from the start for the |
26 |
> benefit of Linux, without making the design compromises that Lennart has |
27 |
> decided suit him. I don't know if forking systemd is easy, but no one has so |
28 |
> far decided to do so. |
29 |
|
30 |
I don't think forking would attract much developers. Writing something |
31 |
new trying to follow "the*nix design principles", but being modern and |
32 |
with the same features (all of them optional, of course) of systemd |
33 |
will have more chances; although I think it will fail because most of |
34 |
the people that can code "better" actually like the systemd design, |
35 |
and would prefer to contribute to it. |
36 |
|
37 |
And if you found enough of this mythical good-coders, good luck |
38 |
defining what it means "the*nix design principles". |
39 |
|
40 |
> Given the title of this thread I fear that those of us who can't code, will |
41 |
> increasingly find our choices becoming limited, because more and more |
42 |
> functionality is hacked inextricably into systemd and friends. It's probably |
43 |
> too early to call if Gentoo will remain one of the few options in Linux that |
44 |
> do not use systemd, but decisions taken upstream (for example initrd for |
45 |
> separate /usr) are affecting some us already. |
46 |
|
47 |
First of all, Gentoo uses systemd if the user so desires (like I do). |
48 |
|
49 |
Secondly, no one has proposed (AFAIK) systemd as the default init |
50 |
system for Gentoo, and I don't think no one will in the short term |
51 |
future. |
52 |
|
53 |
And to finish, the fact is that people are using systemd because it |
54 |
works, the design if good (it can be improved, of course; everything |
55 |
can), and it has attracted a really large flock of talented developers |
56 |
around it. |
57 |
|
58 |
No other option offers any interest for people trying to develop new |
59 |
cool things and design new standards; the only similar (albeit much |
60 |
more limited in scope) alternative was Upstart, and I personally don't |
61 |
think it will be maintained for much longer, except for bugs and |
62 |
security vulnerabilities; it will have no new features. |
63 |
|
64 |
In general the people not wanting to use systemd don't even care about |
65 |
its features; they only want the good old SysV (or OpenRC here in |
66 |
Gentoo), and that nobody touches their systems. |
67 |
|
68 |
Since OpenRC is the default in Gentoo, and I don't think that will |
69 |
change anytime soon, they can have that. |
70 |
|
71 |
Regards. |
72 |
|
73 |
[1] http://lwn.net/Articles/584176/ |
74 |
-- |
75 |
Canek Peláez Valdés |
76 |
Posgrado en Ciencia e Ingeniería de la Computación |
77 |
Universidad Nacional Autónoma de México |