Re: Éric Vyncke's No Objection on draft-ietf-httpbis-cache-header-09: (with COMMENT)

Hi Éric,

Thanks for the comments.  Responses below.

> -- Abstract --
> I am puzzled by the use of "updates it" where the "it" is rather undefined...
> especially as this document 'codifies' it, hence, this is the first time it is
> documented so no need to update it. If I am wrong, perhaps good to add a
> reference to the updated document ?

'it' refers to the immediately preceding noun, 'practice' -- i.e. there's a widespread non-standardised practice. We're not using 'update' here in the sense of 'IETF document update'.

> Also wondering about the use of 'codifies' in a standard track document, i.e.,
> I was expected a 'specifies'. But, as a non-English speaker, the subtle
> differences among the English in different continents probably escape me :-)

'codification' in the sense that it's collecting and restating currently informal practice.

> -- Section 2 --
> 
> Out of curiosity, why do all parameters start with "sf-" ?

Because they're ABNF from Structured Fields (RFC8941).

> How is the IP address specified ? Should RFC 5952 be referenced ?

At this level, the construct is an opaque string; we're not expecting people to validate that it's an IP address or hostname, or use that for any processing. 

> "The Cache-Status header field is only applicable to responses that have been
> generated by an origin server." but how can a cache know whether it connected
> to the 'actual origin' and not another level of CDN ? (possibly a very naive
> question)

This is a good question. I've tried to clarify here:
  https://github.com/httpwg/http-extensions/commit/560389ff2

Is that better?

> -- Section 2.2 --
> Should there be a "other" value to catch up any other cases ?

That would encourage divergence of behaviour in implementations, and reduce the value of the specification. The current values should capture the prominent states in the HTTP caching model; if not, we can add new parameters.

> == NITS ==
> 
> -- Section 2 --
> Just wondering about the capitalized 'List' in 'Its value is a List' when the
> rest of the section uses lowercase 'list'.

That indicates it's a Structured Fields List. The other instances of 'list' are talking about manipulating the underlying data structure.

Cheers,




--
Mark Nottingham   https://www.mnot.net/

Received on Thursday, 5 August 2021 04:14:33 UTC