Hey, John. I'm a little disappointed that you think I think "sensible" is a compliment, and "not sensible" is a pejorative. That comment was directed at Adam, who is on the IESG, hence "not sensible." It just means that we are all a bunch of people who are are more interested in exploring and tooling to do things in ways that work for us than in doing what everybody else does. That's why we are good at developing standards. But we are not our target market. That said, I just did the exercise of having the latest version of Windows 10 download a file with "Content-type: text/plain; charset=utf8", and it doesn't put a BOM on it. Edge displays it correctly; MSIE does not. If you download it with Edge, WordPad and Notepad both screw it up, although the dev editor gets it right. This is really terrible, and Microsoft ought to fix it. There is no excuse for using the iso-8859 encoding when you've been explicitly told the encoding is utf8, and if utf8 is not the default encoding, then when you download a file encoded that way, you have to do what it takes to make sure the file is processed correctly. Windows is the only operating system I tested with this problem. MacOS does fine. Debian does fine. So I think there is actually a serious problem here in need of correcting, but I don't think it's something we have the power to correct, other than in the sense that I hope someone from Microsoft is reading this and either nodding in agreement because they've been trying to get this fixed and haven't been able, or because they see it as an action item. Anyway, I think we should stop arguing about this right now, and the people who have the power to fix the one remaining BOM offender should try to do that. And we should revisit it again in a few years.