Re: Is systemd a benefit or a liability? [Was: btrfs :(]
I have followed on this topic on this list and on other places and I may say I am very concernd over the issue. I believe Craig has summed up the current situation well. A couple of further points, one of the biggest complaints over the development of systemd is the attitude of the main developers, ie them taking little notice of anyone else. The second point is the apparent helter-skelter expansion of facilties in a single system component. The more complex a single item of software becomes the far more difficult it becomes to maintain it bug free. The future of linux with systemd does not bear thinking about. The supporters of systemd do not really make me feel any better, as there is little discussion going on. The pro argument apparently being, its good, the debian developers like it so there. This sort of pro argument goes no where. What is debian going to do if and when the systemd developers take no notice of any changes they propose and this appears likely. Website http://boycottsystemd.org/ may be of interest. Its very likely I have installed my last up date of Debian. I will carry one with the current stable while I assess what alternatives are around. Looks to be three options so far... Gentoo Slackware I have used both of these, the third option in the medium term is to stick with Debian 7.x but gradually update system compnents by building them myself. Note, I already do this to some extent, usually running a later version of the kernel, Xorg and Mesa. This of course does muck up the package managment but the does not bother me much. I could of course make my own deb packages, this __may__ make it easier in the long run. Lindsay Linux since kernel 0.96d, who would have predicted it would end this way? :-(
On Mon, 13 Oct 2014 08:42:58 zlinw@mcmedia.com.au wrote:
I have followed on this topic on this list and on other places and I may say I am very concernd over the issue. I believe Craig has summed up the current situation well.
You can say the same thing about kernels. But all attempts to add micro- kernel features to Linux have been rejected very firmly by everyone who matters.
A couple of further points, one of the biggest complaints over the development of systemd is the attitude of the main developers, ie them taking little notice of anyone else.
Yes, that's a problem. But we have lots of other upstream developers who do similar things. With a project of the scale of Debian we have to deal with all kinds.
The second point is the apparent helter-skelter expansion of facilties in a single system component. The more complex a single item of software becomes the far more difficult it becomes to maintain it bug free. The future of linux with systemd does not bear thinking about.
Systemd is still significantly less complex than the kernel by every measure. Also by most measures it's significantly less complex than libc, Libre Office, and any X based web browser you would want to use nowadays. In terms of the impact of bugs, the kernel is worse than init for security and reliability. Libc is probably similar to init, and web browsers have significant security challenges.
The supporters of systemd do not really make me feel any better, as there is little discussion going on. The pro argument apparently being, its good, the debian developers like it so there. This sort of pro argument goes no where.
Your mistake is to think that we need to convince you. The debate is between developers regarding the best technical options for developing a distribution of Linux. We had that debate and the result was that systemd was chosen.
What is debian going to do if and when the systemd developers take no notice of any changes they propose and this appears likely.
Then we will fork it. Just look at the history of glibc, Xorg, and Libre Office for examples.
I have used both of these, the third option in the medium term is to stick with Debian 7.x but gradually update system compnents by building them myself. Note, I already do this to some extent, usually running a later version of the kernel, Xorg and Mesa. This of course does muck up the package managment but the does not bother me much. I could of course make my own deb packages, this __may__ make it easier in the long run.
If you are going to do that then build your own packages, it's not that difficult. Then you might consider contributing changes to Debian and apply to become a DD, after that your opinion will matter.
Linux since kernel 0.96d, who would have predicted it would end this way?
I've been using Linux since kernel 0.5x. I started before Slackware was available. I've been through the change from BSD init to SysV init, a.out to ELF, and more. -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/
Russell Coker <russell@coker.com.au> writes:
By most measures [systemd is] significantly less complex than libc, Libre Office, and any X based web browser you would want to use nowadays.
That's just an argument for simplifying those other things ;-)
On 13 October 2014 10:01, Russell Coker <russell@coker.com.au> wrote:
Then we will fork it. Just look at the history of glibc, Xorg, and Libre Office for examples.
Doesn't everything use the "nc" browser these days?:-) === cut === brian@aquitard:~$ nc www.google.com.au 80 GET / HTTP/1.0 HTTP/1.0 302 Found Cache-Control: private Content-Type: text/html; charset=UTF-8 Location: http://www.google.com.au/?gfe_rd=cr&ei=RjM7VIa3NMTC8gfGyIGgBw Content-Length: 262 Date: Mon, 13 Oct 2014 02:04:54 GMT Server: GFE/2.0 Alternate-Protocol: 80:quic,p=0.01 <HTML><HEAD><meta http-equiv="content-type" content="text/html;charset=utf-8"> <TITLE>302 Moved</TITLE></HEAD><BODY> <H1>302 Moved</H1> The document has moved <A HREF="http://www.google.com.au/?gfe_rd=cr&ei=RjM7VIa3NMTC8gfGyIGgBw ">here</A>. </BODY></HTML> === cut === All the information you never wanted with a few difficult to remember keystrokes. What could be simpler? We don't need SysV or systemd or anything else, just type the commands in by hand. -- Brian May <brian@microcomaustralia.com.au>
Brian May writes:
brian@aquitard:~$ nc www.google.com.au 80 GET / HTTP/1.0
I see your nc and raise you perl: $ GET Usage: GET [-options] <url>... Of course, that's not part of busybox. One of my RSN-but-really-it'll-never-happen pet projects is to port Debian from coreutils to busybox ;-)
On 13/10/14 08:42, zlinw@mcmedia.com.au wrote:
A couple of further points, one of the biggest complaints over the development of systemd is the attitude of the main developers, ie them taking little notice of anyone else.
The reasonable man adapts himself to the world: the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man. -- George Bernard Shaw
participants (5)
-
Brian May -
Jeremy Visser -
Russell Coker -
trentbuck＠gmail.com -
zlinw＠mcmedia.com.au