RE: Using URIs to identify non-information resources

> What is "http://sdkfsdkfjskldfjksldjf.example.com" the URI of?
 
> How about "http://example.com/dsfsdfsdffdfdf" ?
 
> If I try to visit
>       http://larry1.example.com
> and I get a "301 Moved Permanently" redirect to
>       http://larry2.example.com
> should I still consider ...larry1... to be the URL of your website?

All of these are straightforward.
  - the resource that is connected to at host "sdkfsdkfjsd.....example.com"
    using the HTTP protocol and a path of "/"
  - the resource that is connected to at host "example.com" using the
    HTTP protocol with path "/dsfsdfsdffdfdf"
  - The resource identification  "http://larry1.example.com" does
    does not depend on the behavior of the web server at
"larry1.example.com".
    The URI identifies 'the resource connected to at host larry1.example.com
    using the path "/"'.
    Operationally, what happens will vary, but the activity of connecting
    and getting a response is something that happens after resource
    identification.

I said

> If my web site starts returning "303 See Other", it's still my web site,
> and not me.

and I suppose this wasn't very precise. If my "HTTP server" (not "web site")
starts returning "303 See Other", it still isn't me.

     
> It seems to me you only really know you've got the URI of a web page
> when you get a "200 OK".  These other responses tell you that the URI
> you're working with may be a little bit different.  And this
> realization opens the door to an effective solution to the
> long-standing hash-vs-slash problem.  The solution may not meet
> everyone's intuitions and generalizations perfectly, but it doesn't
> seem to break anything. 

I think the 'hash-vs-slash' problem is a problem of trying to fit
a triangle and square peg into a round hole. Neither fits.

The problem with "303 See other" is that it doesn't fit either.

Larry

Received on Friday, 12 August 2005 07:46:15 UTC