| 1 |
>>>>> On Thu, 27 Oct 2016, Greg KH wrote: |
| 2 |
|
| 3 |
>> Also, I wouldn't completely exclude that we need to change the |
| 4 |
>> wording at some later point. Therefore, we may indeed consider |
| 5 |
>> taking the DCO from the Linux source tree which is distributed |
| 6 |
>> under the GPL-2, instead of the non-free version ("changing it is |
| 7 |
>> not allowed") from developercertificate.org. Their wording is |
| 8 |
>> identical except for the preamble. |
| 9 |
|
| 10 |
> You can't change the text of a license and call it the same thing, |
| 11 |
> which is why that wording is there (same wording is in the GPL), so |
| 12 |
> don't think that by pointing at the one in the kernel source tree |
| 13 |
> that changes anything... |
| 14 |
|
| 15 |
Sure, the text shouldn't be changed without changing the name. I guess |
| 16 |
that's common sense, because otherwise it would be confusing. |
| 17 |
|
| 18 |
> And I would _strongly_ not recomment changing the wording without |
| 19 |
> consulting with a license lawyer, you can mess things up really |
| 20 |
> quickly by changing stuff. |
| 21 |
|
| 22 |
So the DCO was devised by a license lawyer? TBH, I find it less than |
| 23 |
optimal. It is an enumeration with all its items at equal level, but |
| 24 |
its meaning is "I certify ((a || b || c) && d)". That is, structure |
| 25 |
doesn't follow contents there, and at first glance the "or" (or its |
| 26 |
absence) at the end of each item can be easily missed. |
| 27 |
|
| 28 |
Ulrich |
Archives