W3C home > Mailing lists > Public > public-xg-webid@w3.org > January 2012

Proxy and Cache Question

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Mon, 16 Jan 2012 18:59:37 -0500
Message-ID: <4F14B9E9.8090500@openlinksw.com>
To: public-xg-webid@w3.org
On 1/16/12 4:48 PM, Henry Story wrote:
>> >  I don't see how anyone wouldn't understand "mirrored claims" from the above.
>> >  Assertions are made in two places that are semantically equivalent in human and machine discernible ways.
> Ok. So this is starting to be a definition.
> What it the difference between a cache and a proxy? Or are these the same?
You can have a proxy URI, which in our case is a de-referencable URI 
that handles all the nuances associated with Linked Data. You can 
de-reference said URI en route to a graph comprised of other proxy URIs. 
We use this approach to make Linked Data graphs that are guaranteed to 
possess full Linked Data capability.

The same approach applies to bridging of protocols. Basically, what you 
see in our WebID+OpenID bridging which is not different to what existed 
re. openid4.me. It too was using Proxy (or Wrapper) URLs.

As for cache, this comes into play with regards to change-sensitivity 
handling via cache invalidation schemes. In our case, we have default 
schemes as well as those you can override. In SPARQL and the SPARQL 
protocol we allow the use of pragmas for controlling cache etc.. Thus, 
our proxy/wrapper URIs also include cache invalidation smarts that 
leverages what HTTP already delivers.

At their very best, our proxy/wrapper URIs allows anyone (or those 
constrained by ACLs) to progressively construct and exploit Linked Data 
while ultimately enriching the burgeoning Linked Data Open data cloud (a 
massive Lookup data space). Think about the whole world playing a jigsaw 
puzzle game (that never really ends) while continuously producing new 
and useful insights for a variety of profiles, where discovery of these 
insights is increasingly serendipitous due to ever increasing link 
density in the LOD cloud :-)

I am most interested in getting a billion WebIDs in the LOD cloud cache, 
pronto! It can happen if we don't slow ourselves down by being too 
provincial about  WebID re., target publisher, developer, and plumber 
(integrator) profiles.



Kingsley Idehen	
Founder&  CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen

Received on Tuesday, 17 January 2012 00:00:00 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:06:30 UTC