Why Should I Care What Color the Bikeshed Is? (1999)

It's actually very hard to get to the bottom of these things. Back when computers were the sole domain of us geeks and nerds, the culture of downright rudeness was a lot more widespread and tolerated than nowadays. The usual excuse was "content over form", meaning that if I say your patch is stupid because it leaks memory like a sieve, the important thing is that your patch leaks memory, not how you feel about my tone. Which is a reasonable justification when applied to a stupid patch, but turns into a bad excuse when people start calling each other names.

Some of that culture lives on to this very day: we've all read more than one rant by Linus Torvalds where he cusses out someone who tried to submit something stupid to the kernel. Once you get over the fact that the language is uncalled for and that the same message could have been delivered with equal force without cussing, you can actually try to ascertain whether the message itself was deserved or not. In Linus's case, it usually seems deserved.

Which brings me back to Brett Glass, who was indeed called names by PHK and others on the lists. Was that okay? I've already established that I don't think it was, but the real question is whether he "deserved" it, in the sense of whether he really was pissing off everyone on the list with bad ideas and whether there really were people calling for banning him or not.

Like I already said, I don't know the answer, because I really do have better things to do today than wade through vitriol from the 90's. However, if you want to discuss Brett Glass, that's really the question you should be asking. Just because his post on Quora sounds reasonable and polite when compared to nerd-rage that was commonplace on the mailing lists -- the same kind of nerd-rage that would brand anyone as a "colossal douchebag" if used on public forums today -- it doesn't mean his claims were true.

For my part, I would reserve my judgment. On the one hand, it's quite common for an entrenched group to exhibit behavior Glass accused them of. On the other hand, we're dealing with someone who's opposing the Net Neutrality for his personal gain. It's most likely that both sides were at fault, as it usually happens in sordid reality we live in.

It's only bikeshedding when it follows the historical and accepted definition of the term.

Pointing out that a patch has a memory leak doesn't qualify, even if someone randomly calls this critique "bikeshedding".

Pointing out some trivial aspect of the patch, on the other hand, would.

This tends to include things which boil down to personal preference (coding standard whitespacing arguments, emacs vs vim, etc.), naming things (which can easily be renamed later with refactoring tools in static languages), and other minor self-contained decisions which can be easily changed at a later date.

They often don't really understand my use case. They only hear one part of what my team is trying to do, and then act like they know what I need. Yes, your suggestion works for the use case you know about, but not for the 10 other use cases I need to design for.

Second, these considerations were usually already considered while we were planning this work, and there is a reason we decided to do it a different way. I don't have time to re-explain why we chose to do things the way we did to 10 different people who come to me and ask why I didn't choose to do it this other way.

Bikeshedding is when an entire discussion gets bogged down or derailed by people who would rather talk about something that's not super relevant, than say nothing and let other people talk about a topic that's important but they personally have no expertise in. They would rather sidetrack a conversation so that they could be more involved, than sit back and let other people talk about the original topic.

This makes sense for social situations (where what you're talking about is usually far less important than keeping everyone engaged), but makes no sense for situations where the communication really is about solving problems. You can really see this tension on more developer-focused open source mailing lists, where often a small number of people are working really hard on evolving some critical software, but their discussions get drowned out or derailed by precocious teenagers/overconfident CS undergrads/bored retirees who are just killing time on the list as a hobby, and who expend lots of energy trying to get everyone to talk about how terrible the variable naming convention is (or the source file directory structure, or source file naming conventions, or how much they hate this or that open source Internet celebrity person, or what's the deal with airplane food...)

I can totally imagine someone who would dismiss as bikeshedding any comment that they didn't feel like talking about, but that person would be misusing the concept of bikeshedding (and probably being super rude on top of that, as it's an accusation against the presumed bikeshedder as well as a dismissal of what they're saying).