Kamil Paral wrote:
I've spent more than a decade perfecting my email filters and I
have a
setup that works for me very well. I dislike certain aspects of mailing
lists (cross-posting, top-posting, reply-to, etc, which just can't work
well when everyone has to be vigilant all the time to do things right),
but I *like* my existing setup and processes. But that's me, us, the old
timers.
But that is exactly why it is an absurd idea to move away from mailing
lists. Fedora will *lose* all the existing contributors like you or me.
Having said that, my impression is that mailing lists are an already
lost
battle. If you don't have an influx of new contributors, your project is
going to die eventually. Mailing lists are a big hurdle for newcomers.
Young people are not used to it (who still uses mailing lists, in
read-write mode, except for OSS communities?), the lists are difficult to
set up, the user interfaces are bad, there are many peculiarities to be
aware of (top-posting, etc). I don't know if moving away from mailing
lists will make our contributor base grow, but I'm quite certain that
staying with mailing lists will make our contributor base **not** grow.
And our project will slowly decline over time.
I strongly doubt that the mailing list is the main barrier to entry to
Fedora (as opposed to, e.g., packaging guidelines, etc.).
If the issue is that the mailing list is flooding your inbox and you are
unable to filter it, the remedy is simple: Disable mail delivery and use
Gmane NNTP or HyperKitty to read and post to the mailing list. That choice
of preferred technology will go away if we move to a locked-in web platform
such as Discourse. (I know it is FOSS and the data can be exported somehow.
It is still a lock-in for the end user.)
Imagine that you want to contribute to a project and you discover
they're
still using svn, or cvs. There are still some. I personally wouldn't be
bothered, I'd just invest time elsewhere. I value my time.
I do not see the problem. I just need to bring up another UI (Kdesvn or
Cervisia instead of Git-Cola), so what? All I care is that I can update from
upstream and commit/push my changes.
Heck, I still use SVN for *my* personal projects.
And not everything is better with git. SVN basically guarantees a linear
history whereas with git, I have to follow a specific workflow for that
(always pull with rebase, never with the default merge strategy, and for
work branches, always rebase and force-push them rather than merging
master/main into them, then when done fast-forward them to the master/main),
and when working with other people, I usually cannot get them to use it, so
the history becomes a complicated DAG with a mess of merge commits, grrr!
I already have some experience with Discourse and so far it seemed OK
to
me.
I have some experience with Discourse as well and it is just a pain:
* A silly reliance on JavaScript and AJAX for everything instead of server-
side code. In particular, this means it takes several seconds to render a
page on mobile devices such as the PinePhone.
* A constant requirement of the latest&greatest bleeding edge browser, no
support for QtWebEngine LTS branches.
* An absurd assumption that everyone is new to the Internet, leading to lots
of ridiculous gamified spam "achievements" for basic things such as replying
to a thread, with which you get bothered on every single Discourse forum you
(have to) sign up to.
etc.
I do not understand why everybody is moving to this annoying piece of
software and throwing away not only mailing lists, but also web forums using
much better software (i.e., pretty much any other forum software), for it.
But I haven't used it as frequently and in such a volume as
mailing
lists. I also still haven't built my workflow alternative to my email
workflow in there. But I'd be happy to experiment with it. But in order to
get real-world experience, it would be great if this was a shared
experience, where a mass of users really moves and starts using it as a
primary discussion platform. Then we can collectively figure out the
benefits and drawbacks and any workarounds for those drawbacks.
The drawbacks are well known (see above) and the platform is basically
unusable. There is no need to experiment with it to find that out. And
"get[ting] real-world experience" by forcing everyone to use it "as a
primary discussion platform" is just unacceptable.
Kevin Kofler