I was there, I'm personally convinced it's much worse now. Perhaps I'm suffering from rose colored glasses, wouldn't be the first time.
My general impression is that the abuse in the past tended to concentrate in a few well known figures, who often got banned on sight or at least generally carried little to no real influence. Or it concentrated in particular toxic projects which you could just avoid. A little bit of abuse wrangling contributed to project cohesion.
Today it's more like an intermittent firehose, with clueless mobs being weaponized from other platforms. Performances being made in issue trackers and pull requests for offsite audiences, and just a particular depth and relentlessness of harassment that was uncommon decades past.
It's just because there are more people online compared to the 90s. If 1% of people are extremely excessively unpleasant, then with 100 people that's one person. With 1,000 people then that's 10 people, with 10,000 people that's 100 people, etc. There's maybe 10 million developers involved today? So that's 100,000 people.
It's the scale that makes banning the "few well known figures" on sight hard. The method of communication (mailing list vs. GitHub vs. Discord vs. whatever) doesn't really matter all that much, at least not for this particular problem.
I agree that scale is the bigger factor, but the infrastructure matters too: Consider invite only vs writable-by-everyone default. Open source development moved from a lot more invite based in the mid 90s to a strong writable-by-everyone social media model particularly with the adoption of Github vs svn/cvs. But the scale issue is an argument that open source should have been going in the opposite direction. More unpleasant actors means more justification for reducing participation.
I suppose the thing that changed with GitHub is that 1) issue trackers became widely available, and 2) the concept of pull requests (and "drive-by" contributions). Before that, it was basically "email me" for most smaller projects (or maybe an email list on SourceForge).
I suppose "ignore some asshole in your email" is easier, and also hard to pile on comments in a private email box. But beyond that, I'd say having an issue tracker is a net positive, as is the ability to create PRs easily.
I wouldn't say that "social media-izing" is a good way to describe any of that, but I understand what you mean.
The main problem is probably more "brigading" from Twitter/X/Mastodon/Facebook/whatnot.
There were trackers before github for sure-- but like it's a pretty ordinary setup to make issues private to project members until triaged... that sort of thing pretty much kills brigading dead and it's a thing you can't do on github.
There is a continuum between totally private and what github provides, with self hosted tools you could hit a number of different spots on that spectrum. Cynically one reason for the lack of flexibility is that the maximally permissive side is probably better for the expansion of Github's business.
Comments
I was there, I'm personally convinced it's much worse now. Perhaps I'm suffering from rose colored glasses, wouldn't be the first time.
My general impression is that the abuse in the past tended to concentrate in a few well known figures, who often got banned on sight or at least generally carried little to no real influence. Or it concentrated in particular toxic projects which you could just avoid. A little bit of abuse wrangling contributed to project cohesion.
Today it's more like an intermittent firehose, with clueless mobs being weaponized from other platforms. Performances being made in issue trackers and pull requests for offsite audiences, and just a particular depth and relentlessness of harassment that was uncommon decades past.
It's just because there are more people online compared to the 90s. If 1% of people are extremely excessively unpleasant, then with 100 people that's one person. With 1,000 people then that's 10 people, with 10,000 people that's 100 people, etc. There's maybe 10 million developers involved today? So that's 100,000 people.
It's the scale that makes banning the "few well known figures" on sight hard. The method of communication (mailing list vs. GitHub vs. Discord vs. whatever) doesn't really matter all that much, at least not for this particular problem.
I agree that scale is the bigger factor, but the infrastructure matters too: Consider invite only vs writable-by-everyone default. Open source development moved from a lot more invite based in the mid 90s to a strong writable-by-everyone social media model particularly with the adoption of Github vs svn/cvs. But the scale issue is an argument that open source should have been going in the opposite direction. More unpleasant actors means more justification for reducing participation.
I suppose the thing that changed with GitHub is that 1) issue trackers became widely available, and 2) the concept of pull requests (and "drive-by" contributions). Before that, it was basically "email me" for most smaller projects (or maybe an email list on SourceForge).
I suppose "ignore some asshole in your email" is easier, and also hard to pile on comments in a private email box. But beyond that, I'd say having an issue tracker is a net positive, as is the ability to create PRs easily.
I wouldn't say that "social media-izing" is a good way to describe any of that, but I understand what you mean.
The main problem is probably more "brigading" from Twitter/X/Mastodon/Facebook/whatnot.
There were trackers before github for sure-- but like it's a pretty ordinary setup to make issues private to project members until triaged... that sort of thing pretty much kills brigading dead and it's a thing you can't do on github.
There is a continuum between totally private and what github provides, with self hosted tools you could hit a number of different spots on that spectrum. Cynically one reason for the lack of flexibility is that the maximally permissive side is probably better for the expansion of Github's business.