CAW: Was Re: "resolution mechanism"

I would further add that HTTP HEAD/GET is made even more useful with the 
addition of explicit URI resolver services within the response header 
ala the Content-Addressable Web 
(http://onionnetworks.com/caw/draft-onion-caw-02.txt).

This is based on the URI-RES services suggested in RFC 2169, but instead 
of being based purely on convention as is specified on 2169, these 
resolver URIs are made explicit.

For example, with 2169 by itself, you would have to guess that the 
following URI was valid:

http://somehost.com/uri-res/N2R?urn:sha1:RMUVHIRSGUU3VU7FJWRAKW3YWG2S2RFB

...which is really a non-starter since you really have no way of knowing 
which web servers support RFC 2169.

Instead, with the Content-Addressable Web, we add new few new response 
headers:

X-Content-URN:  urn:sha1:RMUVHIRSGUU3VU7FJWRAKW3YWG2S2RFB
X-URI-RES:   http://file.com/pub/file.zip ; N2R ; 
urn:sha1:RMUVHIRSGUU3VU7FJWRAKW3YWG2S2RFB

Since the above uses N2R it is providing the resolution explicitly, if 
you instead use N2L or N2Ls, you would then have a more indirect and 
loosely coupled mechanism for tracking and resolving URIs.


Mark Baker wrote:
> On Fri, Apr 12, 2002 at 01:07:39PM -0400, Simon St.Laurent wrote:
> 
>>Asking for can also be asking about.  Asking about is not necessarily
>>asking for. 
>>
> 
> Ok, let's try and compare apples to apples here...
> 
> 
>>Sure - DDDS lets you query for URI metadata.
>>
>>http://www.ietf.org/html.charters/urn-charter.html
>>
>>"Tell me about this URI..."
>>
> 
> DDDS appears to be analagous to HTTP HEAD; "tell me about this
> resource", whereas HTTP GET does what HEAD does, but also returns
> a representation of the resource itself.
> 
> Does that make better sense?  If so, HTTP does both, whereas DDDS
> only does one.
> 
> MB
> 



-- 
Justin Chapweske, Onion Networks
http://onionnetworks.com/

Received on Friday, 12 April 2002 16:40:53 UTC