Le 01/07/2017 à 10:20, Alessandro Selli a écrit :
> On Sat, 1 Jul 2017 at 00:24:59 +0200
> Didier Kryn <kryn@???> wrote:
>
>> Le 30/06/2017 à 23:30, Vincent Bentley a écrit :
>>> Along time ago, well designed and well behaved software could be
>>> recognised by the number of years passed without changes. I worked at
>>> one place that was proud of release version birthdays.
>> Agreed. That's why I'm always bothered with the fast release pace
>> of some essential pieces of software, with the fastest being the GCC and
>> the Linux kernel.
> Well, this is understandable: both the compiler and the kernel must keep
> up with the frantic rollout of new hardware from many vendors. You don't
> have to upgrade if you don't need support for the laters pieces of silicon
> the industry has churned out the past month. What matters is that older
> releases are still maintained and that they work, of course. From time to
> time support for a new protocol/filesystem is merged, too. This is not the
> case of init systems, as they are not supposed to be sensitive to the
> particular hardware/filesystem they are running on. I wonder why zap feels
> the need of a "regularly updated" init, are there issues in the current
> Devuan init that he feels are not being adequately addressed?
>
New hardware and language evolution are good reasons to
continuously make limited upgrades, but these folks ar continuously
reworking the internals of their machines, like if they were never happy
with what they've done six months before - it's not only bug correction.
They preserve the external API but change all the rest. I tell that
because, a decade ago, I tried to write driver modules but the internal
kernel API was never in sync with the doc, while there was a new edition
of the doc every year; I stopped wasting my time. This might be a reason
why hardware vendors don't provide drivers for Linux: big effort for a
little market.
Didier
::
Re: [DNG] How long should I expect …
