[mythtv-users] Systemd and insanity

[Stephen P. Villano]

On 11/25/13 12:07 PM, Jay Ashworth wrote:
> ----- Original Message -----
>> From: "Stephen P. Villano" <stephen.p.villano at gmail.com>
>>> And that, right there, is my problem with it. It appears to have a
>>> stupidly high Bus Factor; if Lennart gets hit by a truck, *no one*
>>> knows how their machines boot. And since that's my stock in trade
>>> commercially, it makes me really really nervous.
>> So, only one person in the whole wide world knows how to implement or
>> even how systemd works?
> Well, I didn't quite say that.  But, really, what percentage of people who
> use Linux and who are *not* profssional system administrators take the time
> to understand the fiddling details of how the boot sequence actually works?
Well, before I became an SA, *I* did. But, then, I'm a bit... Different.
I enjoy getting into the "nuts and bolts" of nearly everything, from a
household appliance, my car to various operating systems.  :)
> End users, I mean.  Surely the release configuation managers on the distros
> have to understand it at least a little, as do developers of packages that
> need to be started at boot.  I was hoping it was obvious that I considered
> those couple dozen people still to be a pretty small universe vs the number
> of installed Linux systems out there.
Of course, then I recall reiserfs and its inventor. With the fs nearly
being abandoned over one domestic murder case...
>> One person ported systemd to each Linux distro that is using it, with
>> no other developer knowing how that was done?
> It's not completely impossible that that's the case, no; one or two people.
True enough. See the reiserfs bit, though there were more than one or
two people developing it when the "controversy" began. I never *did*
figure out what the brouhaha was about, one man commits a crime, an
entire technology (FS) is tainted?!
>
>> Sorry, rather sounds like the old argument about sendmail.cf.
>> Remember, there were three people who fully comprehended sendmail.cf.
>> One was dead, two were insane...
> And had been driven that way by having to read sendmail.cf.
Same here. Excuse me as a hop around the room, making monkey hoots for a
bit...  ;)
> :-)
>
> But that rather makes my point, Stephen: complexity is bad; unjustified
> complexity more so.  I have yet to see a supportable strategic argument
> that the ends justified the means here, since the means include about 40
> unpaid hours of my time to learn systemd as well as I've learned sysVinit
> in the last decade or two.
>
> Especially since, as several people have suggested, the best approach to
> doing so seems to be UTSL.
>
> Cheers,
> -- jra