| 00:00 | <TabAtkins> | Ideally, yes. |
| 00:09 | <gsnedders> | this is wrt xref maangement? |
| 00:10 | <jyasskin> | Yeah, I don't want to have to define anchors for 'Realm' or 'global object' in every spec that uses the terms. |
| 00:11 | <jyasskin> | I'll still need link-defaults since they don't export any <dfn>s, but it's nicer not to have to find the URLs. |
| 00:11 | gsnedders | really just wants to get rid of Shepherd, or at least most of what it does |
| 00:18 | <jyasskin> | Heh. I just want Bikeshed to do the right thing. The backend is an implementation detail. |
| 06:24 | <annevk> | SimonSapin: yeah, we split the interfaces due to different implementations |
| 08:31 | <MikeSmith> | annevk: wondering if you share some of hsivonen’s views on possible problems with the client-hints header |
| 08:32 | <MikeSmith> | other than what you’ve expressed in https://github.com/whatwg/fetch/pull/258 and wherever |
| 08:51 | <annevk> | MikeSmith: I think content negotiation can work under limited circumstances today |
| 08:52 | <annevk> | MikeSmith: when tightly defined and with interoperable clients |
| 08:53 | <annevk> | MikeSmith: still complicated though, but CDNs and general server knowledge have advanced |
| 09:01 | <annevk> | gsnedders: make it happen |