Lists Home |
Date Index |
Didier PH Martin wrote:
> John said:
> Fair enough. But you can get a Persistent URL from purl.org, and
> you don't even need to worry about mapping it.
[ Aside: the same trick works for IDN FPIs. ]
> Didier replies:
> Yes you are right. The problem I have with PURL.org is that if the site
> is off, the name resolver is dead. This leads to a system breakdown.
But for abstract resources like namespace URIs and
(some) RDF subjects -- which is what we were talking about --
you don't ever need to actually dereference the thing.
XML applications can still process documents with
xmlns:foo = "http://purl.org/foo" even if the purl.org
server is down; it's only a problem if PURL.ORG goes
out of business and auctions off the domain name.
> on the other hand, the URN name resolution based on DNS would took off,
> even if a DNS server is off, the name resolution can still happen. This
> is because DNS includes server redundancy and caching. Said differently
> is more robust and scalable. Maybe DNS URN name resolution didn't
> happened because Microsoft and IBM weren't behind it... with their huge
> money pockets :0
URN schemes haven't taken off because too many people fail
to accept the idea of a universal name that doesn't ever
need to be resolved in order to perform a useful function
(despite the prolific abundance of such names in, e.g.,
XML namespace names and RDF subjects). Every time
someone proposes a non-resolvable URN scheme, it gets
shot down on the grounds that it can't be resolved.