Closed mnot closed 4 years ago
julian.reschke@gmx.de commented:
From 2229:
Apply many editorial suggestions (see #463)
Applied except for the change with respect to the placement of the Considerations for new schemes (separate ticket: #464)
unassigned
to 23
incorporated
new
to closed
@mnot changed summary from Editorial suggestions
to Editorial suggestions 3
2.1 "It uses an extensible, case-insensitive token..." I know what's meant here, but the token itself isn't extensible; it's an extension point. Suggest: :It uses a case-insensitive token as a means to identify..."
2.1 "...based upon a challenge received from the server..." --> "...based upon a challenge received in a response" ("from the server" implies that the next inbound server generated the challenge, which isn't always the case).
2.2 "Note that there can be multiple challenges with the same auth-scheme..." --> "Note that a response can have multiple challenges..."
2.2 "Unless otherwise defined by the authentication scheme,..." --> "Unless specifically allowed by the authentication scheme..."
2.3.1 "... can only be used once per challenge/credentials." --> "...once in a challenge or credential."
2.3.1 needs references to p6 for cache-control directives it talks about.
3.1 "The client MAY repeat the request with a new or replaced Authorisation header field." s/client/user agent/
4.1 'The "Authorization" header field allows a user agent to authenticate itself with a server..." s/server/origin server/
4.3 "Its value consists of credentials containing the authentication information of the user agent for the proxy..." s/user agent/client/
6.2 First paragraph - all instances of 'server' to 'origin server'
6.2 "This is of particular concern when a server hosts resources for multiple parties..." s/server/origin server/
6.2 "...by using a different host name for each party." s/host name/origin/ (or similar; i.e. you can do this with ports too)
Also, the requirements and considerations for registries in our other parts are defined in the IANA Considerations section; here, they're defined in the main document (2.3). Why the difference?
Reported by @mnot, migrated from https://trac.ietf.org/trac/httpbis/ticket/463