[uf-discuss] Fragment identifiers and XFN

[Brian Suda]

On 6/8/07, Ryan King <ryan at technorati.com> wrote:
> It's actually not that simple. A more difficult example:
>
>    http://technorati.com/about/staff.html#tantek_celik
>
> That page has identifying information for a number of people on it.
> XFN doesn't define how to process a part of a page, only an entire page.

--- i've been thinking about this too, without much of a conclusion. I
could see how a page full of people you might want to reference a
single one with the fragment identifier, but that isn't helpful in
other situations and isn't really XFN.

For example:

http://example.org/#vcard

apache would send a 200 HTTP code saying it fetched the index.html,
but you have no idea if it actually contains a fragment called #vcard,
so applications could not rely on web architecture alone, they would
have to parse the page as well.

Whereas,
http://example.org/#foobar is 200
http://example.org/foobar/ might return a 404

I think ryan has a good point, that XFN is a only for the entire page,
so if i own an entire page, i own any fragment you provide on that
page, so that could lead to the conclusion that you could remove any
#foobar fragments when you normalize the URL.

I'm not sure that would apply to any query strings?
http://example.org/users.php?user=briansuda

that could return a unique page with a proper HTTP response code,
because that is evaluated on the server side. But that's another
discussion.

-brian

-- 
brian suda
http://suda.co.uk