I am happy to have discussion such that they are in github issues and
are meaningfully and understandably present on the email lists.
There are two problems.
One is that the current git side tooling does not do a good job of
supporting this. (Some peoples reactions describe it as much worse than
"not a good job".)
The other is that a number of people have objected to the requirement
that it be possible to effectively participate in all WG discussions via
email.
Yours,
Joel
On 1/5/2021 12:32 PM, Kyle Rose wrote:
On Tue, Jan 5, 2021 at 12:20 PM Keith Moore <moore@xxxxxxxxxxxxxxxxxxxx
<mailto:moore@xxxxxxxxxxxxxxxxxxxx>> wrote:
If the new tooling were anywhere nearly objectively better overall,
you might have a point.
Generally speaking, there's no such thing as "objectively better", so
that's a rather high bar to reach. But clearly there are participants
who have found GitHub to be an amazing productivity enhancement.
Why in the world should IETF penalize the vast majority of its
participants in order to favor open source software developers?
Because that's exactly what this is doing. I like open source
software too and definitely want us to be inclusive, but not at the
expense of significantly penalizing other participants.
How is this penalizing participants? You can still engage on the mailing
list in the traditional way. No one is required to use GitHub to
participate in any WG, and chairs should push back in the event an
author implies that a PR is required for a contribution.
Does GitHub make it *easier* for those who know the tools to contribute?
Of course: that's exactly the point. Learn the tools and you too can
benefit.
And I can see why the model of making it easy to submit and manage
text changes, late in a document's development, can make sense for
IETF in general. But git/github is still a really poor interface
for this, and PHB is exactly right that this actually impairs and
splits the discussion.
It does sometimes split discussion, which is where the WG chairs need to
step in and move discussion from an issue to the mailing list when that
happens. If that isn't happening, talk to the chairs or the AD.
We did the experiment. Now it's time to stop the experiment,
collect some results and learn from it.
From where I sit, it's been a tremendous success. Why would we
voluntarily hamstring ourselves by moving back to a less efficient
universal model for contribution?
Kyle
