[lug] discussion/change of topic: new startup tools

[Nate Duehr]

> The author(s) of upstart are doing a re-write right now. It turns out
> that the way they were doing their model didn't work at all for some
> use cases, so they went back and did a redesign and are doing a
> re-write. Once thats done, I expect both Fedora and Ubuntu will start
> moving scripts to native upstart. 

So... big picture-wise here... 

Sun seems to have started this trend with creating a new way to start up
Unix systems in Solaris 10, and then the Linuxes copy-catted it but made
it completely different (as usual)... 

What I've been trying to figure out for a while was... what REAL
benefits does any of this have over having just about everything on
SysV?  

Yeah, there were always the BSD's that did their own non-SysV thing, but
up until this trend started, one could study ONE way to start up Unix
systems and know it real well... now we've just added complexity for
what appears to me to be VERY little real-world gain. 

Isn't that backward on the "progress" scale?  Do we really need to
create new and different ways to confuse newbies that there are now ten
different ways to launch Apache, depending on what distro they use?  

This doesn't seem to be the path to stability or consistency, and
certainly doesn't follow the KISS principal.  Is there something SO
compelling about upstart (or any of the others, now that we're letting
the developers create yet another Unix tower of babel) that is SO USEFUL
it outweighs the goal of simplicity?

Thoughts?  My thought is... yet another example of bored devs who just
kinda do whatever the hell they want, users and sysadmins be damned...
and I won't "blame" Linux for it... conceptually, Sun started this round
of stupidity.

Just being a curmudgeon and asking the question.  I know it'll have
little effect on the herding of the cats that is Linux development... 

--
  Nate Duehr
  nate at natetech.com