James M Snell schrieb:
== Proposal ==
{{{
8.5 Slug: Header
When posting a resource to a collection in order to add a new member, a
client MAY include a Slug request header. This constitutes a request by
This doesn't require RFC2119 terminology. Clients by definition may
include any header they like.
So: "...a client May include a Slug request header."
the client that the URI assigned to the new resource incorporate the
string provided in the value of the Slug header. Server implementations
MAY attempt to comply with the request. The syntax of this header MUST
Again. What's the "MAY" for here? If it's truly optional, no RFC2119
terminology is needed, as the default for unknown headers is to be
ignored anyway.
conform to the isegment-nz construct as defined in Section 2.2 of
[RFC3987]. The value MAY contain characters from character sets other
than [ISO88591] only when encoded according to the rules of [RFC2047].
Again IMHO incorrect usage of RFC2119 terminology. Make it:
"Characters that can not be represented in ISO8859 MUST be encoded using
the rules defined in [RFC2047]".
From a practical point of view, it seems to me that things would be
*much* simpler if we'd specify that the value is an URI (not IRI)
segment, because that would take away the whole dependency on RFC2047.
For example,
POST /myblog/fotes HTTP/1.1
Host: example.org
Content- Type: image/png
Content- Length: nnnn
Slug: a-picture-of-my-house
...binary data...
}}}
I think it would be good if the example indeed would use characters not
in ISO8859.
Best regards, Julian