Re: Improper NMU (Re: NMU for libquota-perl)
On Thu, Sep 05, 2002 at 10:14:40AM -0500, Manoj Srivastava wrote:
> >>"Torsten" == Torsten Landschoff <email@example.com> writes:
> Torsten> You mean we have a better mechanism to not update it. [...]
> Torsten> [...] And I bet Anthony for example does
> Torsten> not have that much time...
> In other words, you do not have either the time, motivation,
> and arguments to convince fellow developers that your idea has
FIGHT FIGHT FIGHT!
Policy's not appropriate for things like this, so that line of argument
is fairly irrelevant. The correct place is the developers-reference,
but that attempts to document consensus, rather than proscribe it.
Which pretty much means the best way to do things is either to pronounce
them from on high, or to discuss them on -devel. Either of which are
perfectly reasonable, although with different properties. Pronouncements
have the benefit that they're efficient -- we have a simple mechanism
for it (the DPL, the release manager, etc), and it doesn't have even the
fairly minimal bureaucratic overhead that policy has. The drawbacks, I
suspect, are fairly obvious to anyone here after a moment's thought. One
of them is that they can be risky: if the pronouncement's a stupid idea,
then we need to recover from that, and we've got plenty of mechanisms
for that too: flamewars on -devel to get the pronouncement revoked,
or countermanded, or if worst comes to worst a general resolution or
And that's great. We should have those flamewars if it was a bad idea. Or
we should have productive discussions on improving it, or how to document
it in the developers-reference, or whatever. Complaining about the contents
of pronouncements is perfectly sensible and reasonable.
Complaining about the mere *existance* of them, however, is daft. If
you're getting the right result -- and if you read the message, I think
you'll find that it did discuss all the issues here, especially advance
notification to the maintainer, and letting active maintainers make the
fix rather than wantonly NMUing -- then more efficiency is a *good* thing.
> I don't know how to avoid that, unless you think we should let
> any changes at all into policy, even when the proposer has not
> offered any convincing rationale for the change.
Letting changes that have had convincing rationale into policy would be
a good start, don't you think? :)
Anthony Towns <firstname.lastname@example.org> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
``If you don't do it now, you'll be one year older when you do.''