| 1 |
On Sat, Jan 26, 2019 at 5:52 PM Rich Freeman <rich0@g.o> wrote: |
| 2 |
|
| 3 |
> On Sat, Jan 26, 2019 at 4:04 PM Kristian Fiskerstrand <k_f@g.o> |
| 4 |
> wrote: |
| 5 |
> > |
| 6 |
> > My personal opinion is we should have a default accepting FSF and OSI |
| 7 |
> > approved free/libre licenses and require acceptance for anything else |
| 8 |
> > though package.license / ACCEPT_LICENSE. |
| 9 |
> |
| 10 |
> From a practical standpoint is this going to block anything used on |
| 11 |
> our stage3s or boot CDs needed for hardware support, such as firmware |
| 12 |
> blobs/etc? I imagine most packages like this would not have |
| 13 |
> FSF/OSI-approved licenses. That includes linux-firmware. |
| 14 |
> |
| 15 |
|
| 16 |
I think the stage3 is already pretty minimal anyway, I'd be curious about |
| 17 |
what it would lose if we did this. |
| 18 |
|
| 19 |
|
| 20 |
> |
| 21 |
> I'm not sure if those are installed by default or how essential they |
| 22 |
> are to actually boot/use any common hardware. |
| 23 |
> |
| 24 |
|
| 25 |
I want to avoid having a singular product here. I think ::gentoo is the |
| 26 |
repo that is the metadistribution and we can basically have defaults there. |
| 27 |
Consumers of ::gentoo are expected to tweak it. I think this is different |
| 28 |
than say, a liveDVD image. The latter we don't expect users to tweak before |
| 29 |
using and we should be trying to support normal use cases. If we need to |
| 30 |
use non-free firmware to do it, I expect us to do that so that users who |
| 31 |
boot the media actually get a working Gentoo install. |
| 32 |
|
| 33 |
So I don't buy an argument that "Gentoo" as a whole has to do a particular |
| 34 |
thing. I expect this discussion is actually more about "the Gentoo repo" |
| 35 |
than about any particular shipped media. Maybe I'm misunderstanding things |
| 36 |
though. |
| 37 |
|
| 38 |
|
| 39 |
> |
| 40 |
> Aside from this, Gentoo has always been more about pragmatism when it |
| 41 |
> comes to licensing. We certainly make it easy to restrict licenses |
| 42 |
> and have a pure-free system, but I'm not sure how painful it would be |
| 43 |
> for users to have this be a default. |
| 44 |
> |
| 45 |
> In particular how likely is this to cause users to end up doing a |
| 46 |
> substantial rebuild 5 minutes after booting their stage3 just to get |
| 47 |
> the system back to a more "practical" state? Granted, bindist |
| 48 |
> probably already causes these sorts of issues but we have no choice |
| 49 |
> there. |
| 50 |
> |
| 51 |
|
| 52 |
I take the Bezos approach here. There are 2 types of decisions: reversible |
| 53 |
and irreversible. This is a reversible decision pretty much, so its low |
| 54 |
risk. If we change the default and the world starts to hate us, we can just |
| 55 |
change it back. |
| 56 |
|
| 57 |
-A |
| 58 |
|
| 59 |
|
| 60 |
> |
| 61 |
> -- |
| 62 |
> Rich |
| 63 |
> |
| 64 |
> |
Archives