1 |
On Fri, 2008-02-01 at 09:58 +0100, Rémi Cardona wrote: |
2 |
> Chris Gianelloni a écrit : |
3 |
> > Is it really that nobody cares or is everybody just being lazy and |
4 |
> > saying "yeah, that's good enough" when I intentionally didn't list lots |
5 |
> > of things that I'd came up with when writing the original list. I |
6 |
> > honestly don't know which it is here. |
7 |
> |
8 |
> Do we know how many people actually use the software on the live media? |
9 |
|
10 |
Yes. Every single person who uses the Installer uses the software on |
11 |
the disk. |
12 |
|
13 |
> Do they care if it's not 100% functional? |
14 |
|
15 |
Yes, since anything non-functional ends up non-functional on their |
16 |
machines. |
17 |
|
18 |
Now, could we focus less on whether checking the software is a valid use |
19 |
of time or not and instead look at *what* we should test within that |
20 |
software? |
21 |
|
22 |
> <personal experience> When I install Gentoo, I'll use which ever live |
23 |
> CD/DVD/USB I can find, if not, I'll download the minimal install CD and |
24 |
> the only pieces of software I'll use is dhcpcd and a web browser to read |
25 |
> the handbook. I'll occasionally use screen and ssh{,d} if the box is too |
26 |
> far from my desk. |
27 |
|
28 |
We don't really care what you use. I mean no offense by this, but it |
29 |
doesn't help *this* conversation any. |
30 |
|
31 |
> So once you release a beta, that's what I'll be trying out first and |
32 |
> foremost. |
33 |
|
34 |
Sure. That's the absolute minimal amount of testing that we normally |
35 |
perform. All we gain is knowing it works on your specific hardware and |
36 |
not much else, since you're planning on testing the parts we test |
37 |
extensively. |
38 |
|
39 |
> But, wouldn't it be interesting to find out the usage patterns for our |
40 |
> users before making a big list of things to check? |
41 |
|
42 |
Not really. The quality should be as perfect as possible on everything, |
43 |
not just the major things. The whole point here is to get community |
44 |
involvement, which means many more eyes so we can cover much more |
45 |
ground. |
46 |
|
47 |
> Maybe not enough users do care to make it block QA and releng's time |
48 |
> when making new releases. |
49 |
|
50 |
We *want* to know what to test. That's the whole point. We've only |
51 |
been testing certain things due to time constraints. If we had a more |
52 |
comprehensive list, it would make a better release. Plain and simple... |
53 |
|
54 |
-- |
55 |
Chris Gianelloni |
56 |
Release Engineering Strategic Lead |
57 |
Games Developer |