1 |
On 04/25/2010 07:36 AM, Ryan Hill wrote: |
2 |
> People make mistakes. |
3 |
|
4 |
Agreed - at work I've often found a quality mindset that is 100% focused |
5 |
on preventing mistakes, and I've found that these kinds of systems are |
6 |
almost equally as focused on preventing them from being fixed (three |
7 |
minutes to fix a bug, three weeks to document and release the fix - then |
8 |
we wonder why the system has hundreds of trivial open bugs with no ROI |
9 |
for fixing them). |
10 |
|
11 |
A good quality system isn't just about preventing mistakes - it needs to |
12 |
be about fixing them too. The system that prevents typos from getting |
13 |
into the tree shouldn't get in the way of those typos being fixed. |
14 |
There needs to be a balance. |
15 |
|
16 |
Scripts running on repository servers don't have a sense of balance, so |
17 |
they aren't the answer. Nor is cutting off hands any time a dev messes |
18 |
up unless it becomes a pattern or there is malicious intent. |
19 |
|
20 |
However, a systematic review process is probably a good thing most of |
21 |
the time, and published policies supporting this are a good thing. |
22 |
|
23 |
Rich |