[9980] in www-talk@info.cern.ch
Re: URL schemes containing +, -, or .
daemon@ATHENA.MIT.EDU (Henrik Frystyk Nielsen)
Wed May 15 11:41:07 1996
To: Larry Masinter <masinter@parc.xerox.com>
Cc: martin@mrrl.lut.ac.uk, www-talk@w3.org
Reply-To: Henrik Frystyk Nielsen <frystyk@w3.org>
Date: Wed, 15 May 1996 11:27:12 -0400
From: Henrik Frystyk Nielsen <frystyk@w3.org>
Resent-From: www-talk@w3.org
Larry Masinter writes:
> > I have an Internet Draft (draft-hamilton-whois-url-00.txt) which
> > specifies a URL format for the WHOIS++ protocol using the scheme
> > "whois." Now, it's been suggested that there really ought to be
> > separate schemes for the old WHOIS protocol, WHOIS++, and RWhois
> > - defined by (respectively) RFCs 954, 1835 and 1714.
>
> Personally, I'd prefer one URL scheme to three. I don't think there
> are clear-cut rules for deciding, though. Are the various WHOIS
> schemes interoperable? And is some of this obsoleted by LDAP?
Regarding whois and whois++ I had a prototype of a whois++ protocol module
in the libwww code some time ago. However, we quickly ran into the problem
that whois and whois++ use the same port number (43 if I recall). As the
protocols are made in such a way that it is impossible to distinguish the
two from each other it is virtually impossible to automate the access.
The rule must be that a URL scheme unambiguously defines an access scheme.
Having two different URL schemes pointing to the same port is not a good
idea unless you have a unique way of identifying them at another level.
Unless this is the case I would strongly recommend using one single URL
scheme.
--
Henrik Frystyk Nielsen, <frystyk@w3.org>
World-Wide Web Consortium, MIT/LCS NE43-356
545 Technology Square, Cambridge MA 02139, USA
