Hello,
The classic Internet protocol for providing per-user data is "finger",
RFC 742 from 1977.
Love it. My first play with redundant/reliable hosting was "fingerhosting",
which
achieved 99.9999% uptime due to tripple servers of 99% each :)
Finger has two drawbacks for this purpose: It is not authenticated nor
encrypted;
Yes, so it is purely there for public data. For such data, it's
better-positioned user data than DNS.
and it is designed to be human-readable, not
machine-readable.
That ought to be good for some degree of privacy ;-) but this is why so many
attempts are made to structure data in DNS but why I prefer LDAP with its large
set of predefined techniques and formats -- and it's openness for DIY specs
that won't clash due to the use of ASN1 OIDs.
I wouldn't mind seeing http://user@domain/ step into this cavity BTW -- HTTP
must be the only protocol on the planet (well, sort of) that does not support
usernames, and we are using this pattern very, very often nowadays.
Given IPv6, putting a unique IP
address per hosted domain isn't so terrible, but having
% finger user(_at_)example(_dot_)com
This would be an operational impossibility I fear. If people need to get an
IPv6 address per user to be able to run finger, then no admin will support it.
"Just use WebFinger", I can hear them say.
WebFinger by the way, is too far up the stack IMHO -- it queries the
.well-known directory on a webserver, fills in a pattern and does a query.
Sounds more like DNS stuff to me, and a good application for
http://user@domain/ -- the other obvious beneficiary being OpenID. This might
call for a SRV record of some kind in the DNS -- or an NAPTR.
(yes, you can finger me for keys to check this message. John convinced me it
the utility 15 years ago.)
Wonderful :) If there were more like you it'd be the IPv6-added-value-showcase
that could help the transport concur the World ;-)
-Rick
_______________________________________________
openpgp mailing list
openpgp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/openpgp