The fewer .well-known usages, the better as far as I'm concerned.  I won't be 
sad to see it go. 

On Wed, Jul 5, 2023, at 16:02, Lucas Pardue wrote:
> Hello QUIC WG,
>
> <Speaking as an editor on the qlog drafts>
>
> The qlog main schema has for a while included a definition about how to 
> retrieve qlogs from a server by accessing a URI that uses a 
> .well-known; see Section 7.2 of draft-ietf-quic-qlog-main-schema  [1]. 
> While this type of thing was considered useful during earlier periods 
> of QUIC interop, research or debugging, it was recommended that 
> production servers not support this access method due to various 
> security and privacy reasons. 
>
> The editors are not aware of this .well-known method being used in 
> actuality. Sharing of qlogs for a range of purposes can be done pretty 
> trivially using alternative out-of-band or custom means. Standardizing 
> the .well-known method would require adding a lot more meat to the 
> bones of its definition, which is work that the editors do not think is 
> worth the effort or possible delay it could impose. Our proposal is to 
> drop this from the spec for now, and let it be picked up in future 
> (perhaps even in a different WG) if people want it. Robin notes that 
> Brian Trammel proposed something along these lines a while back but we 
> don't seem to have tracked it anywhere. So to avoid that again I've 
> created an issue [2] and PR [3].
>
> We think this is an uncontroversial move. However, please speak now if 
> you feel strongly about keeping this in the draft, or are interested in 
> pursue .well-known as a separate item.
>
> Cheers
> Lucas
>
> [1] - 
> https://datatracker.ietf.org/doc/html/draft-ietf-quic-qlog-main-schema#section-7.2
> [2] - https://github.com/quicwg/qlog/issues/313
> [3] - https://github.com/quicwg/qlog/issues/315

Reply via email to