Re: [apps-discuss] Fwd: I-D Action: draft-nottingham-http-browser-hints-01.txt

30.05.2011 1:48, Bjartur Thorlacius wrote:
> The current draft specifies that browser hints are a resource
> identified by the URI "/.well-known/browser-hints" resolved relative
> to an URI scheme, an authority section, and an empty path (assuming I
> understand the draft correctly). I believe this to be harmful, as the
> URI shares the authority section with other URIs, but is in fact named
> by IETF (even though IETF doesn't maintain the resource).
> Have you considered the possibility of passing this information as a
> response to another method than GET, e.g. OPTIONS?
> If creating a new method or reusing OPTIONS isn't an option (no pun
> intended), could a new URI scheme be used, so that the hints are still
> assigned an URI, without polluting the namespace supposedly given to
> domain name registrants, and potentially clashing with existing or
> future URIs.
Bjartur, just my personal opinion.

/.well-known/ has been designed specially to allow everybody and 
everything (meaning applications) to know that something is exactly 
there, and not anywhere else.  Using GET with the definite well-known 
path, denoted as defined in RFC 5785, is probably the easiest way to 
provide a means of access to some particular pre-defined resource in 
HTTP.  I don't understand why we should create a new HTTP method (as you 
proposed in your subsequent message) or a new URI scheme (as you propose 
above) and, in this way, complicate the "protocol", if it may be called so.

What is currently proposed by Mark is fine and needs no further 
improvements (with respect to how the hints are discovered).

Mykyta Yevstifeyev
> _______________________________________________
> apps-discuss mailing list
> apps-discuss@ietf.org
> https://www.ietf.org/mailman/listinfo/apps-discuss
>