| 1 |
Stuart Herbert wrote: |
| 2 |
> Hi, |
| 3 |
> |
| 4 |
> One of the issues that the o.g.o project has brought to a head is the |
| 5 |
> definition of what is "official" and what is not "official" when it |
| 6 |
> comes to Gentoo. The term is already being thrown about in the |
| 7 |
> Project Sunrise thread; I'm sure it'll come up again in future. |
| 8 |
> |
| 9 |
> It's an issue I think we should discuss and find an agreement on. |
| 10 |
> |
| 11 |
> Personally, I think what makes something official or not is 100% down |
| 12 |
> to who does it. I think something is official if it is done by the |
| 13 |
> project (where a project matches the definition in the metastructure |
| 14 |
> project) responsible for whatever we're applying the label "official" |
| 15 |
> to, then that's all that matters. |
| 16 |
|
| 17 |
Its a matter of PR in most cases. Infrastructure has been trying hard to |
| 18 |
make sure any project we host still provides Gentoo with decent PR. Its |
| 19 |
easy for us to say that if its not on Gentoo's servers, then we can't |
| 20 |
fully support it. Its the whole liability thing. (That's why we moved |
| 21 |
all the servers in the rsync.g.o rotation into our control). |
| 22 |
|
| 23 |
User X sees something on *.gentoo.org and assumes that its being |
| 24 |
properly taken care of and managed. If its non-*.gentoo.org, then they |
| 25 |
can be safe to assume its not entirely under the arms of Gentoo. Now, we |
| 26 |
can have experimental stuff on gentoo infra, but the key thing here is |
| 27 |
it needs to be properly maintained and managed. |
| 28 |
|
| 29 |
Say in the case with sunrise, I think a lot of people are concerned with |
| 30 |
the people managing that project won't be able to handle all the |
| 31 |
different types of issues people are worried about. Perhaps its also a |
| 32 |
trust issue also, I'm not sure. |
| 33 |
|
| 34 |
> So (picking something entirely at random for an example), if the Java |
| 35 |
> project had an overlay somewhere (say, on gentooexperimental.org), |
| 36 |
> because it's their overlay, the overlay is "official". Doesn't matter |
| 37 |
> where it is hosted - all that matters is that it is run by the Java |
| 38 |
> project. |
| 39 |
|
| 40 |
Right, and if ge.org gets hacked, its pretty obvious that it wasn't |
| 41 |
officially part of Gentoo anyways. To me "official" means that we (as a |
| 42 |
group of developers) agree to support something in some fashion and |
| 43 |
everyone is held accountable for it since its on Gentoo's central resources. |
| 44 |
|
| 45 |
> Equally (because it is the hot topic of the moment), Project Sunrise's |
| 46 |
> overlay would be "official" because they're a Gentoo project. The way |
| 47 |
> to stop them being "official" is simply to have the Council pass a |
| 48 |
> resolution to shut down the project. |
| 49 |
|
| 50 |
It would have helped if the project had discussed it on ML's *before* |
| 51 |
announcing it to the world and then ignoring all discussion about it. |
| 52 |
I'm pretty sure that the whole attitude they've shown thus far has |
| 53 |
degraded their trust among developers for the project. |
| 54 |
|
| 55 |
The discussion about overlays several months ago specifically was |
| 56 |
against these types of repos being included, yet it somehow got by? |
| 57 |
There was trust involved there that if o.g.o became to being, that it |
| 58 |
would try and keep such repos out. |
| 59 |
|
| 60 |
> I think the other side of the term "official" is clarifying the scope |
| 61 |
> of how far something can be "official". Using the Java project as an |
| 62 |
> example again (sorry guys :), the Java team can put in place |
| 63 |
> "official" policies and procedures for what their team does, but that |
| 64 |
> doesn't make them mandatory for the whole Gentoo project. Other |
| 65 |
> developers remain free to form competitive projects, and put their own |
| 66 |
> "official" policies and procedures in place if they wish. |
| 67 |
|
| 68 |
The trouble here is, those policies don't probably incur more bug |
| 69 |
traffic for *everyone*. There's lots of ways of doing this and most |
| 70 |
people want it done in such a manner to reduce bug traffic, bad PR, and |
| 71 |
an agreed upon policy. |
| 72 |
|
| 73 |
> (I hope I explained that last bit properly. What I'm trying to do is |
| 74 |
> keep in mind the terms of the metastructure document, which explicitly |
| 75 |
> allow for two or more teams to be competing with each other). |
| 76 |
|
| 77 |
I don't think the real argument is a competing team. If it is, what |
| 78 |
teams is it? I'm not sure I understand your point here in relation to |
| 79 |
the current stuff going on. |
| 80 |
|
| 81 |
> What are the alternatives? If a project's activities are not |
| 82 |
> automatically "official", then who gets to decide, and how is that |
| 83 |
> decision made? How can that decision be made fairly, without |
| 84 |
> contradicting the metastructure, and without giving rise to any |
| 85 |
> accusations of 'cabals'? |
| 86 |
|
| 87 |
The decision should be made on our development list for the most part. |
| 88 |
If it seems that most people don't have a problem with it, then it |
| 89 |
should ok to assume that its 'more' official. Now if its discussed and |
| 90 |
several people point out issues with a project, and the project either |
| 91 |
denies or ignores the issues that are brought up, then I would question |
| 92 |
its official status. We're all peers in the same group and we should |
| 93 |
all respect each other's opinions. If such a project cannot work with |
| 94 |
their peers on resolving the issue then it to me the project doesn't |
| 95 |
belong in Gentoo nor be included as official. |
| 96 |
|
| 97 |
-- |
| 98 |
Lance Albertson <ramereth@g.o> |
| 99 |
Gentoo Infrastructure | Operations Manager |
| 100 |
|
| 101 |
--- |
| 102 |
GPG Public Key: <http://www.ramereth.net/lance.asc> |
| 103 |
Key fingerprint: 0423 92F3 544A 1282 5AB1 4D07 416F A15D 27F4 B742 |
| 104 |
|
| 105 |
ramereth/irc.freenode.net |
Archives