[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Gentoo guys starting a fork of udev

On Wed, Nov 14, 2012 at 09:20:05AM -0500, The Wanderer wrote:
> But why is a 30-year-old concept necessarily worse than a new one? Or to put it
> another way, why is it necessary to "bring Linux forward", in cases where what
> is already present is good and works well? (And, taken further: in cases where
> what is already there *isn't* good and/or *doesn't* work well, why is it
> necessary to accept change *in a particular direction*, if that direction has
> problems of its own?)

Because System V Init isn't a good concept. It fails in so many
regards. There is no standardized way for init scripts, it cannot
make sure processes actually run and restart them on demand. It also
lacks mechanisms for ressource control and figuring out dependencies
between service without hardcoding them. It's just a dirty
hack. System V Init was good 20 years ago, but it isn't nowadays.

Automatic dependencies, process watchdogs and ressource control are
something which is incredibly useful to have nowadays, especially on
big machines which are shared among many users (clusters, for

> I've run across a few software projects where it has seemed as if the developers
> were adding new features and removing old ones and changing UIs not because
> there was something wrong with the old, but apparently just because "we're the
> developers, we have to make changes or we're not developing it" - because they
> seemed to think that letting a program sit unchanged is automatically a bad
> thing, no matter how close to perfect-for-its-purpose the program may already
> have been.

True, but as I said, System V Init is not a good concept anymore,
that's why it's being dropped. Apple dropped the old init system with
MacOS X 10.4, why should the Linux world still stick to it in 2012?


Reply to: