On Wed, Jun 1, 2011 at 10:15 PM, J Aaron Farr <[email protected]> wrote: > On Wed, Jun 1, 2011 at 1:41 PM, ant elder <[email protected]> wrote: > >> I had looked already, but ok Tim i went and looked some more but can't >> find so much, do you have links to something specific you remember? >> The main reasons brought up in the past against it which i can find >> are that it might make labs get too comfortable here and that the Labs >> PMC doesn't know how to vet releases. The later of those is hard to >> understand given that all the Labs PMC are ASF members, the first of >> those, if its really an issue, could be dealt with for example by >> giving some so of limit or timeframe a lab can be here doing releases. > > The release issue was settled during the creation of the Labs PMC. > Specifically, you want to search the members@ archive for the > "labs.apache.org" thread. I'll take the risk and expose a bit of the > private discussion here as it may be useful to have it publicly > archived in the future: > > > On Tue, Oct 17, 2006 at 9:49 AM, Stefano Mazzocchi <[email protected]> wrote: >>> - (3) a rule of "no release tarballs" would render most open source >>> projects pretty useless! >> >> If you were to attract a community of end users, maybe, but since labs >> are meant to just distill a pure idea into some rough code that is not a >> concern. >> >> Actually, this is a feature, not a bug: labs will feel suboptimal for >> those who want to use it to abuse the foundation and route around >> incubation. >> >> Think of it as a 'committers repository' with somebody watching over it. >> >> Or think of it as "the development equivalent of the community@ mail list". >> >> For those who want to "try something out and see what others think and >> it's easy to explain the idea in code than in words" don't need >> releases, specific mailing lists or additional community substrate. >> >> They are not doing it to build a community, they are doing it to build a >> seed, that, if starts growing some attention, can become the start of a >> community. >> >> At that point, they can decide to stop working and see if others pick it >> up or push it themselves inside the incubator and properly form a >> community around it. >> >> Labs will exhibit a very high mortality rate compared to the rest of the >> ASF. This is, again, a feature not a bug. >> >> Since there won't be releases, the high mortality cannot reduce the >> quality of our 'long term' commitment on project because users cannot >> depend on it. >> >> On the other hand, the fact that mortality of labs is understood, >> expected and taken into consideration in its own functional rules, >> allows users to be 'wilder' and 'bolder' in their work, removing >> themselves from the burdens of community development and can focus on >> the creativity of showing their ideas with code. > > At the time several of us were very concerned about how much effort, > real or imagined, starting new Apache projects required. The > Foundation was also still transitioning into a more horizontal > organization, spinning out new PMCs and breaking up Jakarta/XML/etc. > Apache Labs was intended to be a broad "sandbox/commons" with a very > low bar of entry and a shared discussion list to encourage > experimenting with ideas and code *in Apache* instead of other hosted > repositories, thereby leveraging and improving the existing trust and > community within the ASF. > > I don't recall much concern that those involved in Labs were somehow > not fit for vetting releases, rather, the entire release process would > introduce more overhead than we wanted in this simple, shared repo. > Moreover, we didn't want Labs to compete with any proper PMCs, > creating conflict when someone brought another project's sandbox to > Labs. > > Personally, I still like the idea of Labs, it's low maintenance and, > at least at one time, served a purpose. >
Ok but there have now been five years to see how it went so some of those concerns can be looked at in a different light. It sounds like from this and what others have said that the rules about releases, committers etc are from a fear abusing the foundation or routing around the Incubator - so the guidelines could be updated to explicitly say that - instead of "Labs cannot do releases." have "Labs cannot do releases in a way they may abuse the foundation or route around the Incubator" - and let the PMC judge the merits of each when its being voted on. ...ant --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
