On 10 May 2013 08:15, Lewis John Mcgibbney <[email protected]>wrote:
> Hi , > > I'm replying to my own mail, however the 1st paragraph of this particular > reply is in response to Peter, his message and to the point(s) he made over > on private@ regarding Git, etc. > > I completely understand about your workflow and the way that you practice. > There is no issue with this. We made the effort to move to Git with the > hope that this would work better for us and it is just a shame that we've > not had more of an opportunity to test that workflow. All hope is not lost > though! > > Once the community understands the basic Git instructions and how easy it is to use Git branches (especially if we can use inline code review in the GitHub/BitBucket style) it should be a smoother process. > I've learnt that in the open source world, no two communities are the same. > Any23 is no exception in this regard. My issue here is not that there are > only one or two of us developing various parts of the library for our own > requirements. My issue is that the reason we develop is to push stable and > improved releases... however we can't satisfy this goal as we can barely > get ANYONE to look at the release candidate. > > Andy replied to the previous thread with a pointer to the list of steps that the Jena project uses, but until I have been through it a few times it will still require me finding a few hours to sit down and focus on it. One step we could do to make it easier may be to enable the rat plugin by default. If the rat plugin was on by default, as annoying as that may turn out to be, we would have picked up missing license headers like the three I missed in my recent Extractor Factory patch before they find their way into the Git master branch. > I don't have a problem persisting with effort within a small community. > We've done it over in Apache Gora and the project has been moving in a > positive direction for the last while now. I've tried to bring as many > positives from that experience over to Any23 as possible, but it seems that > we are really stuck in the mud here. It seems that we relied quite (maybe > too) heavily on our mentors and champions(s), who have now moved on to > other personal and strategic interests. In the resulting dust cloud, the > realization has gradually surfaced (from my perspective anyway), that we > have struggled to sustain and develop the community around Any23 and > ourselves. It now appears that review of a release candidate is not > possible as we need to obtain a level of quorum before we can progress with > release artifacts. The group of individuals that brought Any23 to the > Apache Incubator were (and hopefully are) aware of this process, and I am > optimistic that a nudge of sorts will at least enable us to VOTE and > release Any23 0.8.0. > > After discounting the period where we had svn switched to read-only and git switched on but as read-only, the time frame where contributions have bottomed out is fairly small, especially when you take into account the 3 large feature branches that I have sitting at GitHub which are waiting to be cleaned up and were partially waiting for Sesame-2.7.0 which has been out the door for a few weeks now. Partly at least, the major refactoring of Sesame in 2.7 may have slowed progress on Any23, as I would have liked to have upgraded for 0.8.0 but the timeframes did not work out that way. Overall, I am okay with a community of 2 developers, as long as there are enough people who are well versed in Apache ways to help with release verification. It would also be okay from my perspective if Any23 was merged with SesameTools [1] as a non-Apache project, if it is dropped from Apache. Josh Shinavier and I have been maintaining SesameTools at GitHub--in between his PhD studies and my post-doc/work projects--and it has very similar goals to Any23, being a utility wrapper around OpenRDF Sesame. Cheers, Peter [1] https://github.com/joshsh/sesametools
