Re: SemWeb Non-Starter -- Distributed URI Discovery

Hello,

URIQA FAQ states why a link element (you need to access the
representation first) or a response header (you need an extra network
operation first) is not recommended for providing a link to the
metadata.

Considering that you start with an already made database like Google,
where you already have a cached representation (and maybe the HTTP
headers from the last cache refresh), it would be easy to know which
resources express metadata, rather than doing 8,058,044,651 MGET
requests.

The issue is how much the semantic web and the `regular' web will
overlap at URI level.

I think that for the end-user it has to overlap, such that the SW may
`degrade gracefully'.

Cheers,
-- 
Laurian Gridinoc, www.grapefruit.ro

Received on Monday, 11 April 2005 14:41:35 UTC