I would have no objection to this being published as an Informational RFC,
to document existing practice. But to be completely clear, I was *not* being
sarcastic when I compared it toRFC 3514, because its intended semantics can
be ignored by any web site operator that chooses to do so.
"If you were going to ask me if I wanted to use your 'safe mode', you can assume I'd have said 'yes'" -- it's a pretty hard semantic to ignore, and the cost of doing so is pretty small anyway. My suspicion is that if sites already present the option, there'd be motivation to support the hint (better UX, for one), and no reason at all to ignore it. And for sites that never presented the choice, well, it's irrelevant to them.
It's not really existing practice, either. Yes, content servers providing an option for consumers to receive whatever version/subset of the content the servers consider "safe" (in whatever terms they've couched it) is a thing, but allowing consumers to preempt the question is new, and that's what's on offer. And this proposal does it in the simplest way, which can easily be aligned with the sort of options currently in use (and into the future).
I think there's value in it being a proposed standard.
Matthew Kerwin
http://matthew.kerwin.net.au/
http://matthew.kerwin.net.au/