[Qgis-user] [QGIS-Developer] Announce - migrate our mailing lists to Discourse

[Adam Nielsen]

> I didn't see it coming honestly.  Make it hard for them so only the
> best ones can join us?

Not quite - make asking questions take more effort than doing your own
research, so lazy people are encouraged to find their own answers
before asking.  It's not about excluding people, it's about minimising
the types of questions that could be solved by looking at the
documentation or by doing a web search.

> > In my own experience, if it's too easy for people to ask questions,
> > then they tend to do so before doing their own research.
> 
> Don't you think it is already the case. Moderators everywhere are
> doing this job to remind people to just do their homework before
> asking.

Yes and this is my argument exactly.  Those places make it too easy for
lazy people to ask questions, and so you end up with moderators
spending large amounts of time reminding people of basic things.  In my
opinion, instead of bringing in more moderators, it is better to reduce
the amount of these low-effort questions.

> Would you want to make life harder for those who still want to learn, 
> but just don't get the mailing list stuff?

It is very easy to subscribe to a mailing list.  If you can't figure
that out, I think you will have no hope understanding QGIS.  But it is
not about making it hard or complicated, it is making it so that people
are encouraged to do at least a little work on their own before asking
questions.

> From my experience, stupid questions are either not answered, or very 
> gently refocused on the good way of asking questions, which is also 
> something I like in our communities.

That is true, but it seems this list does not get very many questions
like that.  If we start to get a lot of them, the community's attitude
could change.

> > This is why generally speaking, it's often better to add some
> > hurdles in before people can ask questions, such as figuring out
> > how to subscribe to an e-mail list.
> 
> This is a radical point of view. Another option is to make them pass 
> through helpers that reminds them to search before asking. Something 
> positive, like when you create an issue in gitlab and the UI gently 
> propose you existing similar questions when you start typing.

That is fine too, as long as people cannot ignore it and just keep
typing their simple question.

Look at the Microsoft support forums for example.  They are next to
useless, because it is so easy to post questions.  There are hundreds
of questions asking the same thing, and almost as many "answers" by
people who have no idea what they are talking about, but are trying to
be helpful by suggesting something that worked for them once when they
had a completely different problem.

So although their forums are easily searchable, it is incredibly rare
to get a usable answer from them because the posts are of such low
quality and often outright wrong.

Contrast that to StackExchange.  Yes, the Q&A format is limited,
however by removing low-effort posts and incorrect answers, their
content is consistently high quality and extremely useful.  Many times
I have done a web search and found my answer on a StackExchange site.

Perhaps these are two extremes and the best solution is somewhere in
the middle, but hopefully it at least explains the reasons behind my
arguments.

> > If you still think Discourse is the way to go, I would suggest
> > running it in parallel with the e-mail list for a few months, and
> > compare how many questions get answered there vs here.  
> 
> We can't mirror with both sides being written. If we go this way, we 
> would all have to follow both channels.

I think it would be a mistake to switch entirely in one go.  I am not
saying we run both permanently, just temporarily, to give it a trial
run and see how it goes.  This will give the list users time to try out
Discourse and see how it goes.  Many of the list people will likely set
up Discourse to work like a mailing list anyway, so they won't have to
follow both channels - everything will come into the same inbox.

If Discourse works, then as more and more questions are asked there,
fewer questions will be asked on the lists, and perhaps at some point
they can be turned off and nobody will even notice.

I am reminded of the recent Reddit debacle, where management forced
their changes onto their community of volunteers, and as a result, a
large majority of them left the community entirely.  It does seem like
there is a similar push here, to move everyone to Discourse whether we
like it or not.

And that is fine, I am not paying for the servers, but if having a
helpful community is of value to the project, it is probably best to
take things slowly and give people a chance to adjust, rather than to
dictate what is happening.

> Older ones will stay in mailing lists, Users will try Discourse,
> community is split in half.

Remember the community being split in half, is exactly the same as half
the community leaving because they do not like the change.

> I remind you that the issue is that we do not have any plan at osgeo
> to provide a mailman 3 solution. And new users just don't use mailing 
> lists.

That is fine, I have no problem with it.  I just think it would be wise
to trial the new solution before forcing everyone onto it, and to keep
the possibility open of revisiting Mailman 3 if it turns out that
Discourse is not working.  If you trial Discourse and it ends up
working out really well, well then there's no problem!

> Clearly if Discourse doesn't make it for PSC and user lists, after 
> correct communication and after the transition period, we'll discuss
> it again.

By then you will probably have lost all the "old users" though, so it
may be too late to reverse course.

Cheers,
Adam.