1. Very much so. In Airflow we defined some criteria for becoming
committers and we separated out "code" and "community" contributions to
make it clear.
https://github.com/apache/airflow/blob/master/COMMITTERS.rst#code-contribution
Also what we made clear there that there are no "code/community"
contributors. You can do (and best if you do) both :).

2) I think this is a dangerous path to take. I think one of the important
aspects of Apache projects is how easy and friction-less it is to become a
contributor. Signing an ICLA can be very easily abused by "owners" of the
software to make it not really and truly open-source in spirit. There was
this interesting picture
https://twitter.com/higrys/status/1389979584737779717 which sparked some
short discussion about it.

And while I agree Apache ICLA Is "fine", there are a lot of people who
treat any attempt to ask them to sign ICLA with default "no" or "i am not
going to read all those legal-speak"  or "probably they want to trick me
into something". And for a good reason, because in many cases those ICLAs
might actually have some "interesting" clauses.

I think if we ask people before they are invited to become committers to
sign the ICLA, this goes against this "frictionless" approach and it
immediately raises many questions:

* what is the criteria deciding when we ask ?
* what happens if the contributor refuses?
* what do we answer if the contributor asks why we need it?

I think the current approach where ICLA MUST be signed in order to become a
committer is great. With great powers come great responsibilities, and it
makes perfect sense that the ICLA should be signed then. I see no reason
why we should make the "committer" approval simpler. It brings no benefits
other than a couple of days delay and IMHO, inviting a committer is a thing
that SHOULD look serious and should be involving additional action from the
new committer-to-do. This makes sure that the new committer is aware about
the new powers/responsibilities coming with it.

Side comment: If some projects have a very high bar to become committers,
maybe they should lower the bar rather than ask for ICLA from their
contributors ?

J.


On Sun, May 9, 2021 at 2:58 AM Craig Russell <apache....@gmail.com> wrote:

> Hi,
>
> Looking at How it Works, I think it needs an update to reflect current
> thinking on community.
>
> 1. https://www.apache.org/foundation/how-it-works.html#developers
>
> I believe there is consensus that contributors are not only developers
> with their hands on the code, but:
> people who ask and answer questions on the user and dev lists;
> people who find bugs and report them, with or without writing test cases;
> people who document the projects, including web and "hard copy" documents;
> people who help organize meetups, both in real life and online.
>
> So perhaps we could add a section on "contributors" that covers the other
> categories of non-developer contributors?
>
> 2. I believe that we should ask contributors for an ICLA long before they
> are invited to become committers.
>
> Once a contributor has made several non-trivial contributions to a
> project, I believe that the project should ask them to file an ICLA if they
> have not already done so. This will have these potential benefits:
>
> It will be much easier to make them committers; all it will take is for
> the PMC to hold a successful vote and as soon as they are invited and
> accept, the PMC can simply request their account.
>
> It will give the PMC incentive to communicate with their contributors
> about the value the contributors bring to the projects.  Our increasing use
> of GitHub for development makes this a straightforward exercise. Each PMC
> will have their own criteria for asking for an ICLA, which doubtless will
> be less stringent than committership.
>
> It will clarify the intellectual property issues (provenance) associated
> with the contributions. Some projects have a very high bar for
> committership and all of the contributions prior to formal offers of
> committership are assumed to be given under the terms of the Apache
> License, but we have no formal understanding of this.
>
> Craig
>
> Craig L Russell
> c...@apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@community.apache.org
> For additional commands, e-mail: dev-h...@community.apache.org
>
>

-- 
+48 660 796 129

Reply via email to