On Tue, Aug 2, 2011 at 8:20 AM, C <[email protected]> wrote: > On Tue, Aug 2, 2011 at 14:03, Rob Weir <[email protected]> wrote: >> I'd like us to treat documentation like we do code. Not necessarily >> the same tools, but the same care for provenance, accountability and >> quality, namely: >> >> 1) We welcome "patches" and "contributions" from anyone, but these >> must be first reviewed and approved by a project committer before they >> become part of the documentation set. Any such contributions must be >> made under Apache 2.0 license. >> >> 2) Only project committers have direct write access to the >> documentation. This requires that they first sign the iCLA. >> >> 3) All contributions, whether from the public or from committers and >> tracked/logged, so we can accurately determine who made a given >> change. So no anonymous or pseudonymous patches. A user id that we >> can trace to a real email address is fine. >> >> With code this works by non-committer contributors sending patches >> (diffs) to the mailing list, where they are merged in and reviewed by >> a committer, and then checked into the repository. With >> documentation, using a wiki , we would need a different mechanism for >> achieving this. Luckily there are MediaWiki extensions to enable >> this. >> >> I'd like to preserve the immediate nature of editing on the wiki. >> That is its strength. But we need to find away to also get this under >> project oversight as well. I think we can do both, without too much >> annoyance to contributors. > > This is pretty much JCA regime that was in place under Sun and Oracle. > On the User Doc side, it hindered not encouraged doc contributions... > thus the move to a low entry barrier community Wiki (among other > things that we tried to implement). Accepting patches and fixes via a > bug reporting system is great if you've got the people working the > bugs and managing the input in a timely manner..... otherwise you > simply have a bottleneck in one or two people. The same goes for the > MediaWiki Flagged Revisions (which is installed on the existing OOo > Wiki by the way, just not in use)... without a team of reviewers, the > edits are never approved, and community contributions dry up very > quickly. >
iCLA is not the same as the JCA. You should read it: http://www.apache.org/licenses/icla.txt The OpenOffice JCA says: "Contributor hereby assigns to Sun joint ownership in all worldwide common law and statutory rights associated with the copyrights, copyright application, copyright registration and moral rights in the Contribution to the extent allowable under applicable local laws and copyright conventions." The Apache iCLA says: "You hereby grant to the Foundation and to recipients of software distributed by the Foundation a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare derivative works of, publicly display, publicly perform, sublicense, and distribute Your Contributions and such derivative works." See the key difference? With the iCLA recipients get the same rights as Apache does. There is no asymmetry like there was with the JCA where Sun received special rights. We don't have "a bottleneck in one or two people". All committers are able to review and approve patches. We have (according to Dennis's latest tally) 54 committers on the project right now. How many wiki changes do you think we receive per day? How much time would it take to review and approve a single wiki change? I'd like to see the math that would suggest a bottleneck. Of course, not every committer wants to review documentation. On the other hand, we're not limited to 54 committers. If someone is making a lot of doc changes, and we think they are of high quality, then we elect them to be committers. So in an Apache project, you should never have a review bottleneck. If you do that would be a sign that the PPMC is not doing its job of identifying new committers. > Also you really need to differentiate between Wiki documentation which > is Community developed... and Application help which is/was treated > like the source code (and required a JCA to work on). > All documentation in an Apache project is community-developed. All of it. 100% of it. Every single line. Every character, space, em-dash and en-dash. There is absolutely nothing in an Apache project, code, documentation or website that is not developed by the community. If you are making a distinction between the committers and the contributors and some other "community" then you are making a false distinction. The question is how does the community work within Apache? -Rob <snip>
