I find myself mostly in agreement with Gergő. A reluctance to experiment is
a problem in this movement which prevents meaningful change. The current
state of MediaWiki is such that having discussions on it is very painful.
We can do better.
However, there've been quite a few different experiments with using
Discourse as an alternative to on-wiki discussions over the years. What I'm
left wondering is, what do we expect to learn from this experiment with
Discourse that we didn't learn from the last ones?
Additionally, as an experiment, I think it lacks clear, objective measures
of what would cause the experiment to be branded as either successful or
unsuccessful. These should be defined in advance, along with a plan for how
to measure them, or confirmation bias will means we'll all come away from
this thinking that our pre-conceived notions were proven correct, and we'll
have achieved nothing.
In fact, after I wrote the above, I realised that the exact question of
success metrics was proposed for community input on the talk page for the
get that we like community consultations and all that, but defining a
problem, launching a potential solution, then asking the very people
participating in the experiment what they think the experiment's success
measures should be, strikes me as more of an abandonment of responsibility
than a consultation, as well as invalidating the experiment.
On Wed, 1 Jun 2022 at 17:43, Gergő Tisza <gtisza(a)gmail.com> wrote:
On Wed, Jun 1, 2022 at 12:52 PM Amir Sarabadani
Even if you don't want mediawiki for various
reasons, you can set it up
in Wikimedia Cloud. We already hosted Discourse there for years.
Cloud is 1) not exactly an improvement
in terms of privacy, 2) a drag on human resources as it will take
significant time of an employee or community member (who is likely
unskilled at operating Discourse) to keep the site running. If it seems
likely that the forum will be around for long, it might be worth moving it
to internal hosting (which will be a lot more expensive in relative terms
but still not really significant compared to the Wikimedia movement's
resources, I imagine). In the short term, just buying hosting while we see
how well the new thing works out is a very reasonable approach. Our
community's hostility to experiments is one of the biggest obstacles to
adaptation and addressing long-present problems (such as using discussion
technology that was considered pretty good forty years ago).
Even if you can't host in WMCS for other
reasons, you still can have
internationalized discussions in mediawiki. The Desktop improvements team
does this in mediawiki.org
and while not as great as auto-translate, it works.
No it doesn't, which is why you almost never see multilingual discussions
on meta. It "works" in the same sense that two pieces of stick work as a
lighter: it can be used for the same purpose with sufficient effort, but
that effort is so high that almost no one will use it in practice.
Language barrier is a problem but so is privacy, there is a reason we host
everything onsite. For example, I don't know
the details of how it uses
Google Translate but it is possible we end up sending some data to Google
that are either not anonymized or can be de-anonymized easily. Not to
mention the cloud provider hosting the website having access to everything
and so on. And not to mention auto-translate is not perfect and can cause
all sorts of problems in communication.
While that's a good point and something to consider if we keep Discourse
around, the current reality is that discussions mostly happen on Facebook,
Telegram and Discord, all of which are worse in terms of privacy than a
Discourse site hosted by a contracted organization. These discussions
remind me of the trolley problem a bit - is it really preferable to let
five times more people get run over, just because that way we can wash our
hands afterwards and say we didn't officially approve of either option?
Wikimedia-l mailing list -- wikimedia-l(a)lists.wikimedia.org, guidelines
Public archives at
To unsubscribe send an email to wikimedia-l-leave(a)lists.wikimedia.org