[linux-elitists] How about a fork of Jessie without systemd?

Tomasz Rola rtomek at ceti.pl
Thu May 26 08:48:37 PDT 2016


On Wed, May 25, 2016 at 09:52:00PM -0700, Don Marti wrote:
> begin Tomasz Rola quotation of Wed, May 25, 2016 at 01:00:54AM +0200:
> 
> > Somehow this thread is sliding into minutiae. Perhaps my
> > fault. Anyway, my original attempt was to hint that percentage of
> > users who are going to enjoy systemd's specific functionality is
> > rather small and with such assumption (which could be wrong) the
> > universal adoption of systemd looks strange to me.
> 
> Think about it not from the point of users, but from
> the point of view of developers of server software.
> Does systemd mean that they can spend more time
> on application-specific functionality and less on
> boilerplate service management code?  You're not just
> saving lines of shell in an init script.

There are quite a few OSes out there, some with their unique init
systems. So my guess is, if I were a server developer, introduction of
systemd would mostly mean that for years to come I would have to
support yet another init system. Because there are still servers
running for years without reboot, for example, and many of todays
servers will run with old init system on the grounds that systemd is
not enough tested for security. I definitely would not install systemd
on my hardware just because "oh shiny". I guess I will wait for few
more years and see what happens. Heck, maybe I will even find time to
read systemd source code.

I do not think apache or bind or sendmail will out of a sudden drop
support for *BSD or, say, Open Solaris. I do not think any sane person
in their teams would propose or accept such a move. On the other hand,
even if they did, a replacement can be found or forked. Like, I truly
dislike the current and growing interconnection of gnome and systemd,
to the point that I install xfce when I would have installed gnome
before (the previous objections like memory consumption also helped to
make a switch).

Actually, I believe all software with dev history starting before y2k
had to support multiple OSes or deal with user made patches. And this
is kind of expected, many OSes are here to stay. The only reason to
welcome systemd is to be in a very specific niche (like, Linux-only
software development, certain classes of virtualization with fast
changing use patterns, perhaps some cloud providers etc). Given what
is going on (as I perceive it) I am slowly working to minimise my own
dependency on Linux (includes making choices of softwares and/or
hardwares). And as of containers, despite lots of praise given here
(systemd helps using containers, etc) my specific needs are better
served with full time virtual machines.

So even if I wanted to develop a new software, I would have to make
sure it can run beyound Linux world, in case I would want (or have) to
run it there. In a way, systemd is going to contribute to software
portability, only not by adopting more developers (like you suggest)
but by making them cautious and aware of need to support a number of
choices. Of course, there are probably more and more developers who
only know and care about Linux and being locked in is not a problem
for them nor do they think about it this way. They sound quite like
Windows-only folk, they ignore world outside their own, they might pay
the price in a future like we all pay for our choices (me included, if
I make bad choices).

Times are changing. I can now talk about "locked in" and "Linux" in
same sentence without ever mentioning "Windows".

-- 
Regards,
Tomasz Rola

--
** A C programmer asked whether computer had Buddha's nature.      **
** As the answer, master did "rm -rif" on the programmer's home    **
** directory. And then the C programmer became enlightened...      **
**                                                                 **
** Tomasz Rola          mailto:tomasz_rola at bigfoot.com             **


More information about the linux-elitists mailing list