Re: CAching mechanisms for reducing Network traffic & Latency
Wed, 28 Sep 1994 14:56:10 +0100

NGS == N G Smith <>
DR == Dave Raggett

DR> One idea I would like to see explored is to make a start at handling

NGS> Is the Web now too large for such a radical new idea?

I don't think that the Web is too large to do a considerable amount of work
in getting URNs functional. Although achieving a ubiquitous implementation
might take a while (and probably should, considering the number of issues
still needing to be resolved), there's no reason that individual service
providers couldn't at least have URNs for their own site, providing at least
a partial fix for the "roaming URL" problem.

The work to be done to in getting "site-wide URNs" (ahem) isn't really all
that bad. I know that at least I (and a few others) have made a couple of
baby steps in that direction utilizing HTTP redirection (that is to say, one
attempts to retrieve <URL:>, the
server figures out where to map this to, and issues a redirect to the
client). I'm using the Plexus server, so it's been almost trivial to
implement this at a low level, but it should be doable via CGI programs for
any server.

I personally think that most of the prototyping for URNs for the Web should
be implemented in terms of HTTP redirects instead of spending much effort
modifying the vast installed browser base to do URNs natively. Although
implementing this server-side resolution would kill any large-volume server,
those of us with smaller severs actually have a fair amount of spare CPU
cycles to kill with this kind of experimentation.

Are there any other people who have set up "site-wide URNs"? All's we
really need is two cooperating sites doing server-side URN resolution to get
a large amount of real-world experience in the resolution protocol and the
issues involved. Although I have my own grandiose visions for the
resolution protocol, I imagine that whois++ will be the only protocol we can
standardize upon quickly.

Is anyone working on a perl-based whois++ implementation?

-- | Harvey Mudd College |

"If you've always done it that way, it's probably wrong." -- attributed to Edward Kettering