1 |
>>>>> On Mon, 24 Oct 2016, Rich Freeman wrote: |
2 |
|
3 |
> The end date (which is the one that matters the most) is only |
4 |
> updated when the file is changed. Legally somebody could use an |
5 |
> earlier version of the file when its copyright expired, but they |
6 |
> could only use the latest version when its later copyright expires. |
7 |
|
8 |
> I do tend to agree that we should probably make the start date in |
9 |
> each file depend on when that file was created, but I'm not sure |
10 |
> that legally the start date really matters as much. |
11 |
|
12 |
I disagree. Most of the time, the creation time of an ebuild is |
13 |
meaningless, because they are usually copied from an earlier version |
14 |
of the same package. And I guess that even most ebuilds for new |
15 |
packages aren't written from scratch, but will be based on an existing |
16 |
ebuild or on some template like skel.ebuild. |
17 |
|
18 |
Because of this I also doubt if we could accurately trace authorship |
19 |
of individual ebuilds. Maybe adding an AUTHORS files for the whole |
20 |
tree would make more sense. |
21 |
|
22 |
Ulrich |