Don't export "web user agent".

[jyasskin]

I'm belatedly acting on the agreement in https://github.com/w3ctag/meetings/blob/gh-pages/2025/09-Hong-Kong/09-17-minutes.md#agreement-to-publish-it-as-a-draft-note-and-echidna-it that we can publish this as a Draft Note once it doesn't export "user agent".

Once this is merged, @ylafon can manually publish the first version and get us an Echidna token to autopublish in the future.

---

Preview | Diff

[marcoscaceres]

should we export "web user agent" at all? Might be good to just talk about "user agent" and that's it.

[jyasskin]

#10 indicates that we discussed including "web" in the title in Paris ... at https://github.com/w3ctag/meetings/blob/gh-pages/2025/03-Paris/minutes.md#clarify-the-scope-or-definition-of-what-constitutes-web-user-agent-14 I think.

I'm pretty sure it makes sense to also include it in the definition, for the benefit of other documents that want to be specific about the kind of UA they want to talk about. The lt values ensure that clearly-web documents don't have to use the extra word.

[csarven]

This document does make a distinction between web-UA and UAs in general. https://www.w3.org/TR/UAAG20/#def-web-based2 has something similar "web-based user agent". It can be argued that UUAG's web-based user agent is same as W3C TAG's web user agent but the former is underdefined with respect to the things (such as duties) we wish to be clear about.

If web user agent is defined and uniquely identifiable in this document, then "exporting" is more of a convenience to help with discovery and reuse.

[jyasskin]

"Exporting" is just about whether the term ends up in https://github.com/w3c/webref/blob/main/ed/dfns/ and https://respec.org/xref/?term=user+agent. I believe @marcoscaceres wanted to make sure we got buy-in from the Infra authors before transferring the existing definition from https://infra.spec.whatwg.org/#user-agent.

I read the UUAG's "user agent" as matching our "web user agent" because their definition includes "interaction with web content". Their "web-based user agent" is about UAs implemented in Javascript that can then browse further content on their own. I also filed #33 to continue the discussion about how we describe the category of both-web-and-non-web user agents.

[jyasskin]

I've merged this after discussion in plenary, since none of the comments are about the content of this change. Continue discussing the other points on #33 to inform a future PR.