Ahoj, Dňa Fri, 29 Aug 2014 20:15:21 +0200 lee <lee@yun.yagibdah.de> napísal: > AW <debian.list.tracker@1024bits.com> writes: > > > I'm on the FOSS side. However, systemd is much better than > > sysvinit. > > In which way is systemd better than sysvinit? I am no a systemd fan and i don't like (want nor use) it. But i need to tell, that my understanding is: yes, the systemd is better: - it uses the new kernel properties - it provides more reliably boot in complicated environments - it provides simplest way to work with more amchines at one time - etc > Is it more modular than sysvinit? Does it require less memory? These > are some points mentioned in one of the documents you're referring > to[1]. > > Is it easier to use? Does it have better documentation? Does it > attempt to confine itself to its purpose? Is it more reliable? More > logical? More useful? Or what? IMO, systemd is big step forward, it is a good direction, but the implementation approach is not best. As you mentioned, it is a big blob, despite that it is broken into small parts, because it is terrible to get them to work standalone (one without other). I consider the step from SysV to systemd as step from sequential programming to the OOP. The OOP is there some ten years and it is hard to imagine modern applications, which are not developed by the OOP approach. But there are still huge amount of small things, where OOP is too heavy to use – when i need to resize a directory of the images, iw ill not to develop class a nd method to this, but i will use small script. I see the systemd in similar position. It seems to be good thing, it seems to provide good idea, but for regular users it not comes with something, that SysV cannot achieve. Then it is terrible to find the reason to switch and learn something different for these users. I hope, that in near future there will be developed something, what can enable new kernel properties, but in more modular system (cgmanager as the first step), where one thing can simple work without other. I hope, that the freedesktop will work (contribute) on the standards and not on the particular software. For all these, who don't want to see anti systemd threads here, i can to recall, that the vote for systemd ends in 50:50 (2:2) and only force was used to win (3:2) of the systemd. When i consider this proportion, as the representative of the users, then the one, which used the force, get the responsibility, that other 50% of users will not appreciate this. And by my opinion, there was not considered all consequences of this decision into depth. By this we today can see, that a lot of widely used applications cannot work without (direct or indirect) dependency on the systemd and this is bad! The systemd-shim si not real alternative (at least for me), because some things simple are not working (e.g. X session management for policykit) yet. And it bad too! regards -- Slavko http://slavino.sk
Attachment:
signature.asc
Description: PGP signature