Why can't we setup a meta server sandbox that allows these experimental things to be rapidly activated in the sense of giving each a virtual server slice. That way there is room to play and if something takes off it can then be allocated some serious resources. The ones that die on the vine won't be tying up much of any time or resources since they are virtual anyway.
On 7/12/2011 11:16 AM, foundation-l-request@lists.wikimedia.org wrote:
But now, I feel like we may be able to move back into an era of rapid experimentation, where new projects are more like unmanned 1940s test rockets-- they should be blowing up left and right, as we try to learn from the failed attempts.
I'll go further-- provided we can do so cheaply, I want new projects that are like the ridiculous early failures of flight. [http://www.youtube.com/watch?v=C7OJvv4LG9M]. I want to hear about a new WMF project and it's policy, think "That's crazy-- that's never gonna get off the ground", and indeed, learn something from whether it crashes or whether it actually takes off.
Having an "early flight era" attitude is how we can find something even better than Wikipedia. I agree a lot of ideas are unlikely to work-- but provided the resource usage is sufficiently negligible, let people start making insane flying machine projects, and eventually the wright brothers will show up.