W3C home > Mailing lists > Public > www-talk@w3.org > March to April 2002

Re: "resolution mechanism"

From: Keith Moore <moore@cs.utk.edu>
Date: Fri, 12 Apr 2002 13:50:37 -0400
Message-Id: <200204121750.g3CHobx00874@astro.cs.utk.edu>
To: Mark Baker <distobj@acm.org>
cc: Keith Moore <moore@cs.utk.edu>, www-talk@w3.org
> > > Can you identify anything that doesn't have at least one possible
> > > representation of its current state?
> > 
> > an email address has no state.
> 
> I misspoke.  It isn't necessary that it have any state in order to form
> a representation.  Identity is sufficient.
> 
> When you click on a mailto: URI, the action that is taken is
> equivalent to GET.  

that's a pretty bizarre interpretation of GET, given that any response 
is going to have to be fabricated by something that cannot possibly 
represent the named resource.

> It could even be implemented as such, using an
> intermediary that returns a form for composing an email.  

not in a general sense. such a form could not, for instance, sign the 
message using the sender's private key, nor could it reference the
sender's address book or the sender's message store to aid in composing
the message.  there's no way to implement this functionality on a web 
server without the sender extending to that server a great deal of trust - 
a level which is almost certainly inappropriate unless the user has 
specifically chosen that server to perform that function on his behalf.

of course you can build a MUA out of a web client and server with forms -
but that doesn't mean that such an MUA can serve as an effective 
interface to any random email address.  any message generated by such
an MUA would be indistinguishable from a forgery.

> > a host has state, but GET is not a useful operation for a host either.
> 
> Why not?  I could imagine a representation that included information
> such as who the adminstrator was, information about the machine (OS,
> CPU, etc..).

I think of the 'state' of a host as the contents of its memory, disks,
etc.  (after all, a computer is just a FSM with *lots* of state)
you could define GET for such information, but it's not likely to be 
a useful operation.

if you want to say that any resource has metadata and that it's always
meaningful to ask questions about its metadata, I'd come closer to 
agreeing with you.  OTOH this doesn't really correspond to the GET 
operation as it is deployed in practice.

> > an IP address has no state, but it's useful as a resource name.
> 
> What it identifies, an IP stack, could have a representation.

The address doesn't really identify an IP stack.  At best it identifies
a location in the network. zero or more hosts might be connected to that
location, but there's no way for a GET to collect all of that information.
In many cases the network retains no state about the hosts' locations at 
all - only the hosts know if and where they're connected, and they're not 
necessarily reachable using any kind of stable identifier.

Keith
Received on Friday, 12 April 2002 13:50:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 October 2010 18:14:27 GMT