If they need to punish obscene reports, punish for those, specifically. When they happen. Don't punish a mild, non-obscene report some other time.

edit for your edit: Thanks for finding an earlier clearer example. That comment was unnecessarily personal and abusive. It deserved reprimand or censure. And if the reprimand(s) are responsible for the less-personal tone of the 2011-06-30 comment, compared to the 2010-07-21 comment, they're working to increase the signal-to-emotional-noise ratio. But the controversy of a hard ban for a lesser transgression now introduces new emotional noise.

While we might disagree on the exact mechanics, I think we can agree that any open-source community would be better without such people.

I couldn't agree with that. Imagine what the landscape of software would look like without all those abrasive "bullies" that Mozilla is wringing their hands about: no Theo, no Zed, no Steve, no Richard ... hell, no Linus, for that matter.

So any open-source community would be better without them, eh?

I don't know about other open source communities. Perhaps they thrive on toxic environments full of bullies. But I do know quite a bit about the Mozilla community and I'm quite certain that we are better off without him.

I expect my users to be thirteen-year-old immature brats with no concept of the effort required to write code nor the time constraints which adults generally have to work within, but I think I'm a special case. :3

Open-source community would surely be better without people being bullies. It's not like anyone's creative ability or technical excellence comes from bullying other people.

I prefer rehabilitation to exile, if at all possible.

Quip's like that don't help matters. I prefer peace to war, but it's meaningless.

How do you propose rehabilitating someone who has demonstrated no desire to be rehabilitated within this context?

Also, is it really the job of the developers to fix the bugs in a user rather than focus on the bugs in the product the user is using?

The worst thus-far-linked example of Mlynarik's Bugzilla comments is from July 2010, after which one or more warnings were made. The June 2011 example that got him banned shows him as sarcastic and excessively dramatic, but he refrains from personal insults.

So the idea that he is rehabilitating, after warnings, is one interpretation of the linked reports.

Asa Dotzler has mentioned other quotes, but without dates/links for ontext. Perhaps they were in-person incidents? If so, that would change the calculus a lot (in favor of hair-trigger banning).

But part of a fair "next strike and you're gone" policy is being clear about what a 'strike' is. Was he warned not to use sarcasm, CAPS-YELLING, and excessive emotion? If so, he had fair warning. If instead he was warned not to be uncivil and insulting, then this latest comment didn't prove his incorrigibility.

I understand what you are saying, and it sounds reasonable, but I disagree. Your basically suggesting that in every case of a banning, they need to publicly shine the light on every infraction. They also need to spell out an exhaustive list of "strikes," and beware if they forget to include CAPS-YELLING in the list.

No. This is an adult we are referring to here. We can reasonably expect that he knew exactly what he was doing, and was merely pushing the line as far as he could go, and just happened to cross it. This wasn't the first time; he was a repeated offender. This wasn't the start of the problem. It was the final straw (which, in my experience, is usually a minor straw when compared to the haystack beneath it).

Edit: Thank you for responding with a more clear comment than you original, btw. =)