Re: Let's have a vote!
On 09/26/2014 03:05 PM, Miles Fidelman wrote:
> Stephen Allen wrote:
>> On Fri, Sep 26, 2014 at 11:27:54AM -0400, Miles Fidelman wrote:
>> As has been stated many times here - A GR was tried but not enough
>> seconds to carry on.
>
> Actually, no - I've been following this, and related threads, from the
> beginning - I have not seen anybody actually mention that a GR was
> tried. Do you have a reference?
>
Not in this particular thread, but it was brought up frequently awhile
back when things first started "blowing up" over the systemd decision.
Part of the difficultly of "discussing" systemd issues is that there are
enough bad actors glossing over or ignoring things that people get tired
of pointing it out. It gets really difficult to tell the reasonable
people apart, and can be draining for those who get involved.
In regards to the decision to make systemd the default: The technical
committee was asked to make a decision, and they did. It was not a
person or small group of people who just up and decided to change things
and force it on others.
There are ways to override that decision, you just need enough dd's to
vote that way. However, no one can get enough dd's to vote on even
*having* a vote on the issue. (Memory might be off, but I think you only
need 5 people for that?) Clearly a full vote would be a huge waste of time.
Despite that, there are still several people screaming that the decision
was a takeover, that some dd's are corporate shills, that choice is
being completely taken away, and/or using fancy words like oligarchy.
(Presumably because it's less inflammatory than tyranny?)
My own personal/technical thoughts have gone back and forth on systemd.
However even when I firmly believed that is was the wrong way to go and
would cause major issues down the line, it was still clear to me that
many people were seriously going overboard.
Can we all just calm down a bit?
Writing this already takes a lot for me, so I'm not going to dredge up
the various links myself (they can be found elsewhere in the list or via
search engines).
Debian has a long and very clear document on the site explaining why the
decision went the way it did. Others have posted links to the serious
issues the current init system has, and how they're getting worse.
Systemd is much more than just an init system, so there's a bunch of
other good technical benefits to using it. (Including security and
stability benefits.)
Yes, there were problems with switching in Debian, but people forget 2
things:
1: Debian testing is just that, testing. There have been innumerable
times in the past where having to transition systems has resulted in
temporary breakage. (Like moving to new major versions of
KDE/Gnome/glibc/etc.) Just because you personally haven't experienced
issues running testing doesn't mean it doesn't happen. Granted most of
the time they didn't result in a non-bootable system, but then most of
the time they don't involve stuff crucial to booting.
2: Jessie is nearing it's initial freeze, so there was more of a rush
to get things switched over so proper troubleshooting could happen. The
end result is there was a short time (a week?) where those who didn't
want systemd had to avoid updates until the shim was fixed. (And that
issue wasn't even caused by systemd change, but a kernel change.) Other
issues have popped up, but that's a normal part of the process.
That said, I understand and agree that there are technical downsides as
well. For example, there is a little too much integration within
systemd itself, however that's mostly a result of the maintainer's
attitudes/viewpoints.
And yes, it does worry that the creators/maintainers seem to be somewhat
hostile/clueless. Normally that's an excellent reason to avoid
something. In this case though, there are numerous benefits that
override that concern.
The major reason why I'm not too worried about that last point is that I
have faith in the opensource ecosystem. (Not great phrasing, but best I
could come up with.) Now that Debian is using it to a good extent, along
with most other distributions, systemd has become an important (maybe
even critical?) part of our infrastructure.
We've seen what happens when developers grow uninvolved in the users and
distros needs. Things get forked. Xfree86 and OpenOffice are just a
couple of examples of this. OO in particular showed how distros used a
common set of patches to fix things the developers didn't care about.
The moment an alternative showed up that was willing to just take the
patches, the distros switched over.
Systemd is a decent base with problems, but it *is* fixable provided
there's a steward to allow them. Honestly, I wouldn't be surprised if
within a year or two it was forked and all the distros jumped on board
practically overnight.
Just my thoughts, I'm probably going to seriously regret getting
involved....
- PaulNM
Reply to: