| 1 |
John Davis wrote: |
| 2 |
> Good day everyone: First, I would like to thank all of those who have |
| 3 |
> participated in this conversation, and lambaste those who have |
| 4 |
> decided to start a flame war. In order for *anything* to get figured |
| 5 |
> out, it is the latter that need time to discuss all avenues |
| 6 |
> rationally. |
| 7 |
|
| 8 |
Fair enough. ;) |
| 9 |
|
| 10 |
> |
| 11 |
> Step back for a minute and recognize the ramifications of those |
| 12 |
> numbers. The product that we baby and work on is used by over a |
| 13 |
> _quarter of a million_ people. Over a _quarter of a million_ people |
| 14 |
> rely on our QA procedure for stable packages, our security, our |
| 15 |
> openness. No matter what your political philosophy, this fact should |
| 16 |
> awe you. |
| 17 |
|
| 18 |
I think this is the crux of all our concern, all management and |
| 19 |
organisational decisions revolve around QA, among other primary concerns |
| 20 |
like architecture and so on. |
| 21 |
|
| 22 |
My opinion is that QA procedures can be created without forming a |
| 23 |
government of sorts, or creating a political landscape. There are |
| 24 |
development cultures which succesfully separate the issues surrounding |
| 25 |
architecture, financing and future directions within the distribution |
| 26 |
from the QA process, by placing it on the outskirts of 'governance', and |
| 27 |
I do think we would be wise to do the same. |
| 28 |
|
| 29 |
> So how can me still deliver the same quality product to all of these |
| 30 |
> people? Organization, rules, and voting, to name a few. I referenced |
| 31 |
> Debian because they have an exceptional model that is in place and |
| 32 |
> working. Yes, I know that Debian is not bleeding edge and their |
| 33 |
> releases take forever. Think though, have you ever heard a complaint |
| 34 |
> about Debian's stability, not only in Debian Linux, but in their |
| 35 |
> management as well? |
| 36 |
|
| 37 |
Debian may be stable, but Woody's sheer age really does indicate that |
| 38 |
something is wrong with their development model. I don't imagine |
| 39 |
suggesting a faster paced release model to their core developers would |
| 40 |
be met with much openness, nor would it be put to a vote. FreeBSD and |
| 41 |
RedHat, to name two, have live package build systems (ports/rawhide |
| 42 |
respectively) and cut a release from these every four months on the dot |
| 43 |
with impeccable QA. I don't think I could suggest such a thing to Debian |
| 44 |
developers without being laughed out of the discussion. |
| 45 |
|
| 46 |
Another member of the previous thread mentioned the Linux kernel, if |
| 47 |
briefly. Look at how well it works in its development model. A core |
| 48 |
group of members make decisions as to where the kernel will be in 1, 5, |
| 49 |
10 years, and the rest of the process looks, to the outside world, |
| 50 |
almost entirely haphazard. There is no need for constitutions or |
| 51 |
elections or a legal department, and yet Linux is thriving and growing |
| 52 |
at an extraordinary rate. And it has few enough QA problems that fortune |
| 53 |
100 corporations use it! |
| 54 |
|
| 55 |
It is my contention that the development culture *creates* the product. |
| 56 |
I believe Debian is what it is now because of the way it is managed, and |
| 57 |
Linux is what it is now because of the way *it* is managed. |
| 58 |
|
| 59 |
If I seem staunchly opposed to introducing Debian concepts to Gentoo, |
| 60 |
it's because I am. ;) |
| 61 |
|
| 62 |
My intention though, is not to attempt to close the discussion. If |
| 63 |
through the process an outcome can be reached which seems reasonable to |
| 64 |
all sides, all the better. |
| 65 |
|
| 66 |
Cheers, |
| 67 |
Brad |
| 68 |
|
| 69 |
|
| 70 |
-- |
| 71 |
gentoo-dev@g.o mailing list |
Archives