[OSM-dev] Removing functionality and giving just No as answer

Дмитрий Киселев dmitry.v.kiselev at gmail.com
Fri Feb 24 17:15:53 UTC 2017


>
> Actually carto has had multiple maintainers for some time. Indeed the
> original maintainer, who I assume you are referring to, is not actually
> doing very much these days.
> While osm.org has only one maintainer we have been getting some more
> contributors recently.

Andy has been refactoring the tests to use factories instead of fixtures (a
> big job) and Herve has been updating the emails we send out to start with.


I didn't mean the exact number of contributors, my thought was mainly about
the way how the decisions are made.

The key thing is that you need some mechanism for appointing people to that
> circle of maintainers who get to vote on which things should be included
> and which shouldn't. On most projects that is done by promoting from those
> making useful contributions, so it's hard to move in that direction until
> we can get more people involved.


Going that way, maintainer will fall into approval buble, people who shares
maintainers vision have higher chance to become a "good" committer,
so feedback from users who doesn't share yours vison become smaller and
smaller.

I agree that comments on gh isn't the message board, and not all commits
should be applyed and authours/maintainers vision should be respected,
but now we are in quite a weird state:

We have discussion and voting procedure for tags even for the very specific
tags, like electrical power scemes.
Yes there are many problems with tag voting too, but at least tag author
could get some feedback from broad auditory not only from same minded guys.
But for software, the community have no way to affect pull requests or
roadmap items.



2017-02-24 12:34 GMT-04:00 Tom Hughes <tom at compton.nu>:

> On 24/02/17 16:24, Дмитрий Киселев wrote:
>
> > at this moment, we have some widely used resources such as osm.org
> > and osm carto, started and threated as single persons gh repository.
>
> Actually carto has had multiple maintainers for some time. Indeed the
> original maintainer, who I assume you are referring to, is not actually
> doing very much these days.
>
> While osm.org has only one maintainer we have been getting some more
> contributors recently. Andy has been refactoring the tests to use factories
> instead of fixtures (a big job) and Herve has been updating the emails we
> send out to start with.
>
> And we (osm community) don't have a way to discuss and evaluate changes
>> in collaborative way.
>>
>
> That can certainly work, and I've discussed it with people in the past
> although not everybody has always agreed with the idea.
>
> The key thing is that you need some mechanism for appointing people to
> that circle of maintainers who get to vote on which things should be
> included and which shouldn't. On most projects that is done by promoting
> from those making useful contributions, so it's hard to move in that
> direction until we can get more people involved.
>
> Note that when I say vote here I don't necessarily mean literally as there
> are different models - some projects have a formal vote of some sort for
> each feature (often requiring at least one +1 and no -1 for example) while
> others allow maintainers to decide on their own but with a reversion
> mechanism if somebody objects.
>
> Note that you do need to preselect the voting group, otherwise anybody
> that can drum up enough +1's can get anything in.
>
>
> Tom
>
> --
> Tom Hughes (tom at compton.nu)
> http://compton.nu/
>



-- 
Thank you for your time. Best regards.
Dmitry.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20170224/f0eb59d4/attachment.html>


More information about the dev mailing list