Maxim Kirillov <[EMAIL PROTECTED]> writes:
> On Wed, Sep 21, 2005 at 07:14:53PM +0100, Andrew M. Bishop wrote:
> > Max Kirillov <[EMAIL PROTECTED]> writes:
> >> HTTP/1.1 302
> >> Location: <URL>#<fragment>
> >>
> >> and wwwofle performs the 'wwwofle -fetch' command, it
> >> recursively requests document <URL>%23<fragment> instead
> >> of <URL>
> >
> > It is not defined in HTTP specification to have a header
> > with a fragment identifier. <...> The HTTP protocol does
> > not handle fragment identifiers anywhere, it is only
> > browsers that understand what they are and they are only
> > valid for HTML type documents.
>
> Let's look further:
> http://www.w3.org/TR/html4/struct/links.html#edef-A:
> ---------------------
> href = URI [->http://www.w3.org/TR/html4/types.html#type-uri]
> ---------------------
>
> http://www.w3.org/TR/html4/types.html#type-uri:
> ---------------------
> This specification uses the term URI as defined in [URI] (see also [RFC1630]).
> ---------------------
>
> The [URI] link leads to RFC2396, wich explicitly says that
> "As such, it [fragment] is not part of a URI, but is often
> used in conjunction with a URI."
>
> So, we can state that href arrtibute cannot contain fragment
> also. But this is absurdly.
>
> I tend to think that this is a flaw in standards. So, there
> is no reason for this flaw not to appear in other places,
> including the HTTP specification.
Yes, there are flaws in standards, even RFCs, are not without problems
in some places. Your argument is that an error in one standard means
that there could be errors in another, this much is true. But it does
not mean that there is the same error in more than one standard.
The HTTP specification is for the transport of data, not the display
of formatting of it, not the links between documents. The HTML
specification is for the display of data, the formatting and links
between documents and not the transport of data. A fragment is only
useful for HTML and as such it is only refered to in the HTML
specification. There is no need to mention it in the HTTP
specification because it has nothing to do with HTTP. A fragment
identifier is useless for a plain text URL or an image or a sound or a
video, the fragment is part of HTML and not part of the URL.
If there is a mistake in the HTTP RFC I would say that the mistake is
to not explicitly exclude fragments from the Location header like they
are from the Referer header (or to mention them as excluded in all
URLs).
--
Andrew.
----------------------------------------------------------------------
Andrew M. Bishop [EMAIL PROTECTED]
http://www.gedanken.demon.co.uk/
WWWOFFLE users page:
http://www.gedanken.demon.co.uk/wwwoffle/version-2.8/user.html
