[svlug] Devuan

Rick Moen rick at linuxmafia.com
Sun Jul 10 23:35:53 PDT 2016


Quoting Steve Litt (slitt at troubleshooters.com):

[udev:]

> The preceding is a perfect example. In a supposed effort to get that
> last, dumbest user, they create an albatross. Which wouldn't be all
> that bad, except they make sure their albatross is absolutely
> incompatible with the easy manual interchangeable part building methods
> we've been using since the Unix founders made two and three letter
> commands to make it easy to pipe them all together to make a 1 line
> application.

I have a confession to make.  I get easily driven into doubt (that I
understand the other party) when suddenly I see nothing but
highest-level abstractions.  I'm sure that it can makes me seem obtuse, 
and I feel bad about that.

In the above passage, I'm honestly not at all sure what you're referring
to.  In fact, it makes me uncertain whether you're speaking about udev
at all, because I cannot tell what the phrase 'incompatible with...
easy manual interchangeable part building methods' would mean.

That would kinda-sorta make sense as an ideological talking point _if_
we had been talking about systemd, but systemd wasn't the subject of
discussion, udev was.

Remember, at the time in question, when Greg K-H had newly created udev,
the systemd people had not yet taking over maintenance of udev and
pointedly refused to lift a finger to help anyone build udev without
also building systemd.

Anyway, were you just mistakenly switching topics back to your favourite
monomania, which was _not_ what we were talking about?  If not, please
explain, because I honestly don't understand what you wrote as a
reference to udev.



> The package manager-foo you discuss in the preceding referred document
> can vanish any time....

How?  And also why?  Some evil conspiracy with magic powers and cryptic
and unclear motives?

This suggestion is IMO transrational in two ways at the same time:  It
blows right past any concern with logical connection concerning

1. How?, and
2. why?

I daresay I have a pretty good understanding of both the technical and
polico-bureaucratic aspects of (respectively) Debian-as-architecture and
Debian-as-project, and I'm utterly mystified about either the how or
the why -- as concerns udev, in particular.


> ...and indeed encounters hostility from many in the Debian
> community...

Um, and when will I encounter that hostility on account of my publishing
that Web page?  (Not to mention why.)

That page has been now up for quite a few months, and is linked from the
front of the without-systemd.org Web site, so it's not exactly been
hiding.

I don't expect to have 'hostility' visited upon me by anyone for having
published that page.  I'm tempted to stake a $10 wager on that, but
there are obvious process problems including your phrase '[person] in
the Debian community' being undefined.

So, forgive an old cynic, but I cannot help suspecting there is no
specific semantic content in what you said.


> starting with this doozy from Paul Tagliamonte, who
> apparently had the power to move Thorsten Glaser's request that
> sysvinit continue as an alternative:
> 
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=727708

You know, I actually _really_ don't have time for this, but I'll let
myself get suckered into this, once.

Bug was filed against an abstract pseudo-package calld 'tech-ctte',
i.e., the Debian Technical Committee -- which is of course not a
software package but rather a group of volunteers.  My understanding
(which may be incorrect, as I'm not a Debian developer) is that this was
_slightly_ irregular, a bit of grandstanding humbug, but not _totally_ out
of bounds per se.  The 'bug' requested the Committee decide on a default
init (meaning technically an init system) for the project.  Thorsten
immediately made the arguable comment that this really was a matter for
a General Resolution, not the Technical Committee.

What you claim is 'apparent' is not correct at all.  Paul Tagliamonte,
to the best of my understanding, did not, and does not, have any
particular 'power' that any other Debian developer doesn't also have.

Anyway, much bureaucratic passive-aggressive maneouvering followed, down
long, long, tedious pages of dogfighting.

At this point... good God, man.  I'm only about 1% down this utter waste
of time, and I already know that what you said makes no procedural sense
whatsoever, and I fear that it's going to take me hours and hours to
read this benighted thing, if I am to have any hope of not losing the
scent. 

And you didn't actually point out what to look for, so this is really
hopeless.  (I believe I did slog through the damned thing three years
ago, but had purged its details from memory as mostly time-wasting
bullshit.)

> Notice how that thread resembles a pro-systemd kangaroo court.

No, actually, in slogging through that irritating mess until my eyes
crossed, I saw mostly a number of developers getting pissy over the
notion of other people telling them what they had to do and citing what
amounted to ideological reasons.  

Which is what volunteers do when they're already doing a lot of work and
someone comes in and tries to tell them they should be obliged to do
more and different work than what they signed up for.  There's a widely
felt urge to tell the someone to smeg off.

I'm not sure I want to spend the next n=several hours reading the rest
of that, especially when your starting assumptions were wrong and your
conclusion looks like a non-sequitur.

I personally think the bug asked the wrong question.

It should, IMO, have asked 'Should Debian continue to use GNOME as its
default desktop, or should it switch to (say) LXDE?'  (TBH, that
particular DE would have been an awkward choice in 2013, because it was
just starting to morph into LXQt.)

_Me_, I'd have _really_ wanted it to say 'Should Debian continue to use
GNOME as its default desktop, or should it instead default to the
Blackbox WM with _no_ DE?'  (Any number of other WMs would be equally
suitable.)  But I fear that defaulting to no DE would have been
considered too radical by many in 2013, much as I think the Debian
Project ought to be realistic and not act like anyone's going to take it
seriously as a novice distribution.


> And then later in the thread the Debian Dev crybabies bemoaned the extra
> work of maintaining both a sysvinit run script (which they've been
> doing for years) and a systemd unit file (which is advertised as being
> dirt simple). 

You know what?  They didn't actually owe you fsck-all, for starters.
I think the whole discussion was misbegotten, and debated the wrong
point entirely, mind you.  But if I'd seen you call them 'crybabies'
because they didn't want to officially commit to doing exactly what you
want, I'd have told you to smeg off, if not used harsher language and
firmer measures to get you out of my hair.

But it's all a tempest in a teapot, anyway:  I've explained to you why
that is, several times, but that rolls off your back every time, and you
ignore what I carefully unpacked about actual software process, and go
straight back to prepared talking points as if I hadn't bothered.

This is an extremely poor use of your time, and of mine.  And moreover,
I have an extremely strong suspicion that you're not even talking to me
at all: This is you doing a 'witnessing', and you're just using me as a
prop for a ritual stump speech.




> They decided not to guarantee continuing sysvinit. 

Let's cut the bullshit.  $100 says I will continue to be able to use
OpenRC as the set of init scripts in Debian 9 'Stretch' using
substantially the same methods described on
http://linuxmafia.com/faq/Debian/openrc-conversion.html for
substantially the same use-cases mentioned.

Level odds, winner pays loser upon release of Debian 9 'Stretch'.

I might be wrong, but strongly suspect I am not, and am willing to take
your money.



> Yes it does. For now. But there were two votes, and both votes
> give the "Debian Devs" carte-blanche to sabotage, either by not caring
> or by malice, non-systemd Debian installations.

I cannot match this description with the historical record at
https://www.debian.org/vote/ .  I see one General Resolution (entitled
'init system coupling') with a number of proposed amendments (four) that 
ended up being chosen among via ranked-choice alternative (Condorcet)
voting.  The winning alternative amendment (#4) was one that turned the GR
text into a statement that circumstances did not actually require a
General Resolution at all.  That having carried, then the electorate
approved that statement via vote.  

That does not equate to what you said.  It equates to 'We, a majority of
the voting members, don't think we need to vote on this today.'

Try to use electoral gamesmanship to strongarm developers in that method
was seen, IMO, as Ian Jackson pulling a 'let's you and him fight' move,
and the developers resenting him trying that and said 'no, go away'.

You prefer to see shadowy conspiracies, that's your privilege, but
frankly to me you sound more than a little loony when you do that.

Now, OK, Steve, we're done.

Don't try again to use me as a prop for a 'witnessing', please.  I would
probably, in the name of decorum, just tell you 'No, go away' -- pretty
much the way a strong majority of Debian developers told Ian Jackson the
same thing about his series of 'init system coupling' ploys.



> "They that can give up essential liberty to purchase a little temporary
> safety, deserve neither liberty nor safety."

And protect your precious bodily fluids!




More information about the svlug mailing list