This thread is just one more in a long tradition of similar posts. It invariably goes like this:
New community member shows up, offers their opinions on julia, often with some praise and with some recognition of things we can do better, sometimes accompanied with some confidently presented statements that some other community members think are factually incorrect or misleading.
Various community members start showing up, attempting correct various misconceptions, or to suggest workarounds for the pain points.
The original poster attempts to respond to everyone, and often in their haste makes more statements that more community members feel the need to comment on. Often, I think that having to respond to so many separate people can lead to a sort of defensiveness in the original poster.
Tone and politeness start breaking down, and everyone starts to see what they believe is evidence of ‘bad faith’ on the other side.
I think the ‘slow mode’ is a reasonable thing to try to stop this dynamic from happening. It seemed to work at first yesterday until things flamed up again today.
@dariush-bahrami and @SadeghPouriyan I really don’t think this was moderator action against you, so much as trying to stop the conversation from going off the rails in a way that similar conversations have in the past. Unfortunately, that attempt seems to have not succeeded and now things are even further off the rails.
For what it’s worth, I still think the slow mode is a good idea, but perhaps there’s something we can do to improve the messaging, or another way we can stop the sort of dogpiling that leads to this stuff before it blows up.
Every single post I made and action I took in that thread was in an attempt to keep it focused and productive, precisely because they can and do often devolve quickly, especially when they gather over 100 comments in 24 hours.
What do I mean by devolve? It’s not casting aspersions on any single participant, it’s that folks have strong feelings and they can end up arguing about things that don’t fit within our organizing principle — a place for civil discussions about Julia — and the arguments themselves can end up being far more hostile than any action I could take as a moderator. (Ah, thanks @Mason, you’ve described it well as I was composing this). So yes, even in #offtopic we hold fast to that.
I earnestly do hope you stick around long enough to get to know us better. As moderators, we use an extraordinarily light touch, sometimes to our own detriment. I strongly favor thread splits over closures. Clearly, we’re letting this discussion continue here for now. Nobody is being censored. In the case of the thread in question, I did want to split off the off-topic moderation bits, but they were very interwoven with the discussion so thoroughly I didn’t see a path forward there. Again, this is a community resource — threads you start aren’t so much “your threads” as they are forming a part of our community. So closing that thread isn’t targeting you.
One unfortunate side-effect of the slow down mode (besides just being interpreted as hostile) is that it meant that folks were more likely to combine multiple responses into a single post, which made thread-splitting challenging/impossible since Discourse doesn’t support splitting a single post. It also limited editing, which is something that I wasn’t aware of, but kindof makes sense to prevent folks from having a conversation-through-edits.
I believe one of the reasons that the slow down mode it might be perceived as hostile is precisely for the third point outlined by @Mason.
So even if it is somewhat preventing a overwhelming wave of statements from the community members, it also might make the OP feel that is being prevented from respond to everyone and get frustrated about it.
Maybe, a call for regular posters or community members to try to refrain from adding too many statement even if they are trying to clarify non-general claims made at some point and we take more time as a community to respond. A call for self-imposed slow down (or impulse control as @helgee said).
Not sure if that would work well enough though, but we may try .
One option might be a slowdown mode that applies to everyone except the OP, to limit the “pile-on” effect. But there is no perfect purely technical solution here.
Indeed, OP may be the person who most needs to post. They started a topic, the discussion diverged, and now they may be unhappy to see muzzled/delayed without resolution. One more post would at least give them the chance to vent or get out the last word.
Another possibility is One Last Word for everybody. The topic is listed as closing soon, allowing one final post for each person. I know that sometimes I just want to explain myself one last time, if I feel I’ve been misunderstood. I have that need even if nobody is listening.
I don’t envy the moderators, who must make decisions under tense circumstances, where almost any decision will make some people unhappy.
These kinds of discussions do appear quite regularly, as @Mason said in his excellent comment above. It seems that they tend to begin in good faith on both sides but can get heated quickly at which point any attempt to mediate the discussion hurts someone’s feelings.
This makes me think that maybe it would work better to treat these discussions differently from the beginning. Maybe there should be a specific Discourse category “Language critique / feedback” and new posts in this topic would have to conform to some sort of template (think issue templates on github) that would require the author to answer things like “Was my critique addressed in What’s bad about Julia talk” and other questions that would require the poster to learn what is reasonable to expect from the language and what is being worked on and prioritized. I can imagine multiple types of such templates for topics like “Workflow issues”, “Performance issues”, “Feature parity with X”…
One thing we could try is to fork the thread as soon as possible if some of the points raised are actually things that can be solved, or discussed at least, as specific technical issues (The long time of VS code to load as one example of that, maybe there is a bug there to be addressed, or some non trivial configuration option leading to that, and it got lost).
When the thread mixes concrete issues with general subjective impresions they tend to go bad.
Personally, I don’t mind this perception that much; not because I think it is true, but rather because I think it raises the barrier to these kind posts.
Surely a lot of things can be improved about Julia and the associated tooling, but
finding about these things usually takes at least a month, if not several, of intensive usage,
a lot of them are already known and have open issues,
in 90% of problems that usually come up in posts like the one linked above can usually be remedied by reading the manual, doing a bit of research (eg comparing IDEs), or asking here.
Finally, Julia users (and especially developers) have a lot to be satisfied about: a great programing language.
.