Hi,

I see another point, how to conserve the history of what has been done in github, I mean, is there a way to apply a patch if the parent commit of the patch hasn't got the same hash in the Apache whiteboard repo than the github one ?

-Fred

-----Message d'origine----- From: Greg Reddin
Sent: Thursday, April 04, 2013 8:59 PM
To: dev@flex.apache.org
Subject: Re: [DISCUSS] How do we want to handle Whiteboard?

On Thu, Apr 4, 2013 at 12:44 PM, Om <bigosma...@gmail.com> wrote:

Personally, I like the option of working on GitHub for the whiteboard.  In
a way it levels the playing field of committer vs. non-committer.  This
could foster more community involvement.


I like the idea of using Github for all the reasons mentioned. But I
think there are three issues that need some discussion before we go forward
with it.

First, how will we ensure that we are developing software in community? If
everything is worked in our repo or submitted patches, then I can follow
the development by subscribing to the commits list. If it's at Github, then
I have to know about it and I have to subscribe to an external feed for
updates. So, for it to work in the Apache Way, the dev list needs to know
where and how to watch it or contribute to it. This seems doable, but it's
not automatic. One thing we want to avoid is one or more people working on
something externally for a long period of time, then bringing it to the
project as a large chunk.

The second issue is ensuring provenance. If a committer commits something
or a contributor submits a patch in Jira we know where it came from and who
contributed it. Does a Github whiteboard give us the same traceability of
contributions? Granted, it's possible for someone to contribute something
they don't have rights to even in the current model, but would we still
have the traceability at Github?

Thirdly, how will we identify folks who should be nominated as committers?
We'd need to know all the places at Github to keep track of and we'd need a
way to be made aware of someone's contributions.

Again, I think these are doable, but I think we need to consider how and
what.

Greg

Reply via email to