Normative status of "5. The User-Agent String" is unclear

[gsnedders]

What is the issue with the Compatibility Standard?

https://compat.spec.whatwg.org/#ua-string-section seems to mostly describe the status quo, per its description:

This section serves as a descriptive record of the User-Agent patterns found in the so-called major web browsers

It would probably be good to explicitly mark this as informative? (e.g., via Bikeshed's informative classes)

I mean the alternative is to make this fully normative, and require implementations to use specific user agent strings (presumably tied to the navigator compatibility mode), but that seems like it would be a much more substantive change than what I believe the current intention is?

[miketaylr]

the alternative is to make this fully normative, and require implementations to use specific user agent strings (presumably tied to the navigator compatibility mode),

I kind of like this idea. If anyone is interested in discussing at TPAC next week, let's do that (otherwise I'll just cook up a PR while pretending to pay attention in some session).

[karlcow]

let's chat.

[miketaylr]

TBD if this should should integrate with https://fetch.spec.whatwg.org/#default-user-agent-value

[miketaylr]

OK, discussed with @karlcow today and we decided the simplest thing to do for now is mark this as non-normative. I think having this be one day normative would be useful, but it's going to be more work and we'll probably want to add some additional machinery to fetch to at least allow for per-request user-agent values (given the existence of quirks and UA overrides), and we'll have to modify things here to account for embedders like Opera, which are probably using a navigator compatibility mode of Chrome, but don't send the Chrome-UA.