W3C home > Mailing lists > Public > www-tag@w3.org > January 2008

Re: [httpRange-14] What is an Information Resource?

From: John Black <JohnBlack@kashori.com>
Date: Tue, 8 Jan 2008 06:23:06 -0500
Message-ID: <052d01c851e8$de5f9e10$6601a8c0@KASHORI001>
To: "Pat Hayes" <phayes@ihmc.us>, <noah_mendelsohn@us.ibm.com>
Cc: "Mikael Nilsson" <mikael@nilsson.name>, "Sean B. Palmer" <sean@miscoranda.com>, <www-tag@w3.org>, <www-tag-request@w3.org>

on Mon. Jan. 07, 2008 at 7:39 PM, Noah Mendelsohn wrote:
> (with apologies for the long vacation-induced delay in replying)
> Pat Hayes wrote:
>> [Noah Mendelsohn wrote:]
>> > That content type licenses the receiver to interpret the octets as
> much
>> >more than characters, but as representing a tree of elements with some
>> >semantics.
>> Right, but note, REPRESENTING the tree. Not actually BEING the tree.
> Let's say I have in mind a tree.  It is the tree that has a node labeled R
> as its root, A as a left child, and B as a right.
>         R
>         |
>       -----
>       |   |
>       A   B
> I mint http://noahmendelsohn.com/trees/sampletree as a URI and as the
> authority I decide that it identifies the tree above.  Are you saying I've
> made a mistake?
> Let's say I further invent a media type, say tree/preorder, I define a
> lisp-like syntax in which every node is a triple of its label, it's left
> child and its right, set out as Unicode characters, space separated, in
> parens.  So, the above tree can be "represented" using that media type as:
>        (R (A () ()) (B () ()))
> It doesn't much matter what representation I choose.  All that matters is
> that it's a media type that can represent trees.
> If you do an HTTP get to my URI, I return a 200 with that media type and
> body.
> The URI identifies the tree.  It's information in the Shannon sense.

You talk as if information is a substance, but it is not. I believe Shannon 
means information to describe the *difference* between what the sender and 
the receiver know. Both parties are needed - it is senseless to talk of one 
side alone as "being" information. Critically, the receiver must *not* know 
what state the sender is in before the message. A message only contains 
information when it informs the receiver.  HTTP 1.1 already recognizes this 
in the following status code definition:
"10.3.5 304 Not modified

If ... the document has not been modified, the server SHOULD respond with 
this status code. The 304 response MUST NOT contain a message-body ..."

For these reasons, IMO, following Shannon, we may perhaps talk about 
"information representations" or messages, but not "information resources", 
since whether the resource holds any information for the receiver depends on 
whether or not the receiver already knows what state the resource is in or 

>  Of
> all the possible trees, I can use a message to help you tell which one I
> have in mind.  Or, if you like, of all possible information resources,
> only some of which are trees, the combination of the 200, the media type
> tree/preorder, and the LISP-like string above tell you which information
> resource I have in mind, I.e. the abstract tree.  Anything broken with
> that?

Well, only that if I already have a copy of the tree, and it has not been 
modified, you should have sent me "304 Not Modified" alone, rather than 
"200" plus the tree, since there is no information in that resource for me.

John Black

> Noah
> P.S. I currently own noahmendelsohn.com but for the record I prefer not to
> actually assign the sample URI above.  You'll get a 404 if you try to
> access it.
> --------------------------------------
> Noah Mendelsohn
> IBM Corporation
> One Rogers Street
> Cambridge, MA 02142
> 1-617-693-4036
> --------------------------------------
Received on Tuesday, 8 January 2008 11:26:14 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:55 UTC