What I found quite derailing upon reading the very first comment was that it raises two non-issues:
- that the proposal should generalise to more keywords than just `type` (it does!)
- that the proposed change could somehow make typos valid programs (it does not; .type must be surrounded by parens (.type), and such a typo would likely trigger a syntax error *even if* -XOverloadedRecordDot was active)
Reading such troll-ish posts (even if the author did not mean to troll) often triggers a strong urge in me to reply to correct these perceived misconceptions.
Soon I'm not the only one replying. The troll (by perceived function, not by self-declaration) keeps on fueling the discussion with ever new contentious material, at which point the discussion has been successfully derailed.
Everyone participating in the discussion *feels* like they are helping, but in reality they are sadly just providing more fuel.
I do not know enough about moderation practices to emphatically suggest a solution.
Of course it helps if the committee itself does not engage with trolls, but there are many other people with a GitHub account who might still engage (and they do!).
In the present case, the proposal author engaged with the troll as well.
That highlights an important issue: The proposal author is supposed to defend their proposal against critique, and rebut any which is invalid.
In other words: *it might seem like it is the job* of the proposal author to engage with trolls.
(Of course, ultimately only critique from the committee needs addressing, but I think it's "good practice" to rebut early.)
If I was a proposal author and the troll accepted a rebuttal as an invitation for more inflammatory discussion that went un-moderated, I would be upset about the experience.
I think that is what happened here. Curiously, I don't see a direct violation of any HF Guideline as I interpret them, but to me it feels like the whole discussion was started by the troll in bad faith.