Yep; very true. I realised that I was hardly able to articulate for myself what my issues were.
So I’ve started on a bit of a “what’s good about Discourse?” search because I know that a lot of people like it, and coming from where its strong points are seems like a more productive thing than trying to identify the dissonance with whatever mental model I’m carrying over from my experience elsewhere.
IOW no more comments about how someone’s MUA is messing up threading. 
Since everyone uses the same tool there isn’t a need to teach people who do not have an email-heavy life/workflow how to use email effectively and make it work with any of the major browser-based email clients that so many people use these days.
Being able to reply to everyone who has commented since you last read a thread in a single response I find can really help when trying to get across a narrative. (And for those of you who don’t know, you can highlight text in any reply in a topic and have it end up in your current reply, and you can do that multiple times in a single reply.)
The fact that you can go back and edit your posts is really nice, especially if you accidentally were a tad rude and you want to tone something down. It even lets admins fix spelling mistakes.
Now I know you can use HTML emails, but I have seen enough people complain when they get such emails that I’m calling this out as a separate thing. I mean I can’t even imagine doing an email as nicely formatted as this post is. This also includes fixing titles so it’s easier to know what a topic is truly about.
When a topic veers off-course or splits into a new topic, good luck trying to get people to do a split and actually follow it. With Discourse an admin and forcibly split a conversation into a new topic and that helps keep things on-topic in a discussion. (See this topic as an example .)
It is so much easier to get people to flag misbehaving folks here than it is on a mailing list. Convincing people to email admins or the Conduct WG is really hard. Compare that to here will there’s a flag icon sitting right there on every post that people can use which is a much lower barrier to doing the right thing and bringing bad actors to the attention of the admins.
And I mean so much easier (just imagine I have a ton of emphasis on how so much better this is to get the scale appropriately ). For instance, accounts are in a single place so if we have an issue we can take care of the person and have it ripple through the whole instance; compare that to the mailing lists on mail.python.org where each individual list has to maintain their own suspensions and blocks and so bad actors can continue to operate in other places.
We get notified when someone has been ignored by five or more people as an early warning that someone might be stirring up trouble. That’s really handy compared to people maintaining private kill lists in email where that information never reaches the admins.
Suspensions can be timed. When we hand out e.g. 3 month suspension on a mailing list I have to personally remember to turn off the suspension or have the suspended person reach out to me personally to undo it. Here on Discourse it’s built into the tool, so the suspension undoes itself automatically.
Spam is automatically managed. On mailing lists the solution is to set all new subscribers to moderated to make sure that their initial email is legit and then clear the moderation flag. Here not only is it harder to sign up for an account to do the spamming, but the system detects when someone posts the same link too often.
IOW I’m totally willing to be an admin on this Discourse instance while I constantly contemplate giving up being an admin for mailing lists as it’s just so much more of a hassle.
I like that Discourse is a hybrid between a listserv and a forum, even allowing people to reply to discussions via email. The email digest is also a convenient way to keep up with the latest, and popular, conversations.
I like having Markdown formatting. I like having a “like” option.
There are a number of other things I like in a qualified way (“I like X but…”). I won’t mention those as the qualifications are important to me and this thread is supposed to be about what people like, not what they dislike.
Thanks, @brettcannon, for launching (forking?) this thread. For wider visibility, I’ve tweeted to suggest that people contribute here https://twitter.com/tjguk/status/1207970105461805057
This is a definite plus for admins. Although I’ve now stepped down, I was moderator/list owner on a dozen Python mailman lists for several years, and the points you make are definitely valid.
This is something traditional e-mail cannot deliver and it brings me some anxiety: I hate typos when I see them, I hate grammar errors when I make them, I want to be as clear as possible in my communication so that other second-language speakers (like me) can understand what I mean.
I choose how I view the posts, how big the font is, what color scheme I want. But there’s still rich formatting by the author. In e-mail we have a few people who keep using 76-character-formatted plaintext, other users that use HTML e-mail (like me).
from discourse import consistent_python_formatting
def look(how: str, nice: str) -> str:
"""Try this, e-mail!"""
return consistent_python_formatting(how, nice)
In HTML e-mail, some customize font sizes to be either tiny which makes it very annoying to read long messages formatted this way. Yet others choose to bump their font size making it look hilarious when content is quoted. Unless I specifically try to be obnoxious on Discourse, this does not happen
Tagging users, quotes that let me go back to the original quoted post and back quickly. Links to external websites which have nice click counters on them.
Examples:
py said:
“Less toxic (so far)”
Can you give a recent example of toxic behaviour on the Python-Dev or
Python-Ideas mailing lists?
The most recent example I recall is the PEP about the := operator – that was so toxic we lost our BDFL.
PEP 572 was accepted in July 2018, so that’s well over a year ago,
closer to 18 months.
And, as I understand it, the toxic behaviour was mostly off-list, on
social media:
If I’ve been informed correctly, there is at least one person who has
been banned in the last few weeks, so people ought to have no problem at
all pointing to recent examples of bad behaviour, and shouldn’t have to
go all the way back to July 2018.
Banned for behavior on Discourse, or the mailing lists?
I don’t follow -Ideas like I once did, and a year and a half isn’t that long – especially for an event that had such a profound impact.
@steven.daprano I appreciate the follow-up. However, I refrain from continuing that point on this thread as it is not something I wish to focus on and may detract from the main topic.
Other things I like about Discourse:
From my perspective and experience so far on both platforms (Python MLs and Discuss), the vast majority of interactions have been highly positive and constructive.
That being said, I don’t think either platform is more or less toxic than the other, but Discourse does have significantly improved moderation features compared to the mailing lists. It might not affect the quantity of toxic individuals, but it certainly can help to reduce their impact on the community.