Re: Spec Updated - http-request & vocabulary

On 5/10/07, Norman Walsh <ndw@nwalsh.com> wrote:
>
> Under http-request, I find:
>
>   When the request is formulated, the step and/or protocol
>   implementation may add headers as necessary to either complete the
>   request or as appropriate for the content specified (e.g. transfer
>   encodings). A user of this step is guaranteed that their requested
>   headers and content will be sent with the exception of any conflicts
>   with protocol-related headers. If the user of the step requests a
>   header value (e.g. content-type) that conflicts with a value the step
>   and/or protocol implementation must set, the step will fail.
>
> But this is much too vague. I don't understand, either as a user or an
> implementor, what headers might be inconflict and when.



For example, if you set the Content-Transfer-Encoding header to
something for which the content can't be encoded, the step should
fail.  I can't see how we can enumerate all of them.

Of course, this is a huge interoperability issue.

I think we must state that setting certain headers could cause
conflict with proper generation of the request.

We could say "the right thing will happen" or we could say "the step may
fail".  We have to say something about this.


> And I suggest you replace the penultimate word "will" with one of the
> RFC2119 words.


Yes, I will.



-- 
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics