W3C home > Mailing lists > Public > www-tag@w3.org > August 2002

Re: URI denumerability

From: Tim Bray <tbray@textuality.com>
Date: Wed, 28 Aug 2002 20:35:50 -0700
Message-ID: <3D6D9696.5020109@textuality.com>
To: "Roy T. Fielding" <fielding@apache.org>
Cc: Norman Walsh <Norman.Walsh@Sun.COM>, www-tag@w3.org

Roy T. Fielding wrote:

> 
> Right.  Any given URI is finite.  However, there is no finite limit
> on the length of a URI.  Therefore, given any list of URI one could
> construct a URI not on that list ...
> 
> In short, that footnote would be wrong even if it were stated correctly.
> We should just delete it and all memory of the concept.

No and yes.  The set of URIs is clearly denumerable.  Let's assume that 
each character of the URI could contain a full-range Unicode character, 
and make an infinite list of all URIs lexically ordered as follows:

0x0
0x01
0x02
...
0x10FFFF
0x0, 0x0
0x0, 0x1
...
0x0, 0x10FFFF
0x1, 0x0
...

Given any URI, I can easily compute its integer index in this list. 
Given any integer, I can easily compute the URI at that index in the 
list.  There are no URIs that are not in the list.  The set is 
denumerable.

Yes, delete the damn footnote. -Tim
Received on Wednesday, 28 August 2002 23:35:50 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:55:42 GMT