[p2p-hackers] Re: [decentralization] The Content-Addressable Web

Mark Baker distobj at acm.org
Thu Oct 25 14:39:01 UTC 2001


> : The N2* conventions run completely against the architecture of the Web.
> : URIs are resource identifiers.  URNs are one kind of URI.  How many
> : URIs does a resource need?
> 
> N2L doesn't provide URIs, it provides URLs.

URLs are URIs, so N2L does provide them.

>  A resource needs at least
> as many URLs has it has addressible locations.

URLs are names.  The only thing that makes them a locator is a
convention that converts them to an IP address, TCP port, local name
tuple.  But that mapping isn't authoritative.  Quite the opposite
in fact.  For example, chances are that you likely have a cached
version of http://www.yahoo.com in your browser cache.  What is
the URL for that document?  It's http://www.yahoo.com.

Is "Tony Kimball" a locator?  Of course not, it's a name, right?
So what if we defined a convention that said that you could do a
DNS lookup on tony.kimball.person, open a connection to port 81,
do a GET, and see his homepage.  Would that make "Tony Kimball"
a locator?

MB
-- 
Mark Baker, CSO, Planetfred.
Ottawa, Ontario, CANADA.
mbaker at planetfred.com



More information about the P2p-hackers mailing list