[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Proposal for collaborative maintenance of packages

On Fri, 2005-12-30 at 07:45 +0000, Neil Williams wrote:
> On Friday 30 December 2005 7:11 am, skaller wrote:

> Until the NM backlog is cleared, I'd recommend applying now if you see any 
> prospect of this changing in the next year. It could be at least 8 months 
> before you get assigned an application manager and over a year before you 
> become a DD.

I'm just curious why it takes that long .. that worse
than a government department :)

> Only once the initial packaging had been checked by a sponsor AND with some 
> kind of stall built-in before it gets into the main pool! 

Sure. My idea is to streamline getting things into the input end of
the pipeline. I'm not trying to get them past quality checks,
contrarily, I'm advocating letting lesser mortals than a DD actually
commission such checks that can be done automatically: a higher
load on the hardware, but more, and more timely, feedback to
the maintainers which hopefully will actually increase quality.

> Maybe your request could be better suited to *exclusive* experimental uploads. 
> i.e. slightly more access to experimental than to unstable. There is no 
> automatic path to unstable from experimental.

Yes, that may be -- this is where my crude idea is better developed
by the real experts.

> > They would have more authority than 'none' but much less
> > than a DD. Exactly how much of what I don't know: enough
> > to actually put the package in place and trigger an autobuild,
> experimental won't necessarily do that.

I don't mean uploading triggers the build, rather that it is
possible to request a build: indeed the two things seem fairly
independent: often a change of a build tool requires a build,
and sometimes you just want to commit source for inspection
without building.

> > not enough to push the packaging into a place it might end up
> > in the stable distro.
> Uck. That sounds like you want a release-critical bug filed against your 
> upload the minute it hits the archive.

That may be one way to implement it: but I can't say what the
best way is.

> IMHO, sponsors are more than sufficient for this. What is REALLY necessary 
> here is:
> 1. Quicker transit through the NM process
> 2. More sponsors

I think we're asking for two things, both of which I would hope
have the same effect -- you're saying lets get more sponsors working,
and I'm saying lets get them working more efficiently.

> For most practical purposes, you'd at least need to be running Debian locally 
> rather than Ubuntu. 

Debian was my first choice -- but when I first built my box 
I couldn't get it to work, my ISP didn't mirror AMD64 binaries, 
and I had a very  limited bandwidth quota to work with. 

Ubuntu built from  one CD image, and it handled my new 
fangled motherboard  out of the box. Also some of the 
core Ubuntu development team (Canonical Employees) happen  
to be active in my local Linux users group :)

> You must have an interest in packaging or you wouldn't be here, 

Yes, of course.

> if only that 
> your software needs to be packaged to find those users. 

That is part of it for sure, but by no means everything.
I also have an interest in building an desktop for dumb
users (like me!) that is patently better than Windows,
so there aren't any arguments that it isn't feasible to use
Linux in the workplace.

This is because politically I would like to see the world 
managed by open processes .. and by far the most important 
processes are the software tools we trust to do the menial work: 
whether it be banking, networking, communications, culture, 
food distribution, manufacturing, voting, or whatever.

So yes, for me Debian is important. When enlightened
people such as those in Peru *mandate* open source software
for all government agencies .. the world is taking a large
step forward towards peaceful coexistence and progress,
and away from greed, power, and authoritarianism.

> It may not be your 
> primary focus, but it does need to be considered.

Yes, certainly. I'm a supporter.

John Skaller <skaller at users dot sf dot net>
Felix, successor to C++: http://felix.sf.net

Reply to: