Re: Moving Ambient Light Tests, Re: Agenda - Distributed Meeting 15 August 2012

On Wednesday, 15 August 2012 at 00:57, Tobie Langel wrote:

> On 8/14/12 11:16 PM, "Marcos Caceres" <w3c@marcosc.com (mailto:w3c@marcosc.com)> wrote:
>  
> > Also, how do we co-ordinate this? Is the W3C Github account or project
> > already set up? Or do individuals just control their own repos and just
> > link to them from the specs (obviously, I don't see that working well
> > personally to help build communityŠ unless someone already popular does
> > it)?
>  
>  
>  
> I think the best option would be to have projects hosted under the
> umbrella of the W3C (github) organization, so they would all sit at
> github.com/w3c (http://github.com/w3c). W3C staff would be able to create Github teams with R+W
> access to specific repos. Repos would all be public so everyone would have
> read access.

Awesome, I see this is all set up already!  Thanks for the pointer. I'll bug the appropriate people to get access and move stuff there.  
  
>  
> How repos are organized and/or name is probably best left to individual
> groups to figure out, though I think matching w3.org/TR/ (http://w3.org/TR/) could prove
> invaluable.

Agreed.  
>  
> > Another question: would the existing Web Apps test suite creation process
> > [1] work for a general community in the wild, or is it too process heavy?
> > How do you invasion test contributions will work (i.e., shared
> > project/contributors, integration control into main test branch,
> > code/test quality control/guidelines, etc.)?
>  
>  
>  
> I don't think much of it would need to change. Inclusion of new test case
> in the test suite could be done through pull requests, burdening the
> puller (a member of the team with r+w access to the repo) to verify that
> the requester has signed the CLA. In practice a high number of
> contributions come from a small number of contributors, which considerably
> lightens this burden. Node.js, which is the second most followed
> repository on github[1], and the seventh most forked one[2], handles this
> process manually; I've yet to hear complains about it.
>  

Makes sense. Thanks for the explanations! That's great! :)

Received on Wednesday, 15 August 2012 00:22:02 UTC