Re: New Internet-Draft: finger URL

Paul Hoffman (ietf-lists@proper.com)
Tue, 14 Feb 1995 17:52:06 -0700


Message-Id: <v02110107ab66fd39fecf@[165.227.40.19]>
Date: Tue, 14 Feb 1995 17:52:06 -0700
To: miked@ncd.com (Michael A. Dolan), uri@bunyip.com
From: ietf-lists@proper.com (Paul Hoffman)
Subject: Re: New Internet-Draft: finger URL

>>     finger:<request>
>
>Doesn't this presume a local fingerd...

No, it assumes a local finger-enabled client. All URLs assume a client that
can speak in the protocol of the Internet service defined.

>...and/or a hard-wired fingerd host/port ?

Yes, it assumes hard-wired finger port. RFC 1288 assumes a single port, and
I've never seen anyone fingering on any alternative port.

>I kinda liked Reed's last draft.......

I would have liked to have seen that. I looked through the URI mailing list
archive and didn't see anything.

>>finger://hostname/[W/][username][hostlist]

I see a few problems with this that I posted in an earlier message. If the
typical user sees the sentence "For more information, finger
nasanews@space.mit.edu", they would most likely form the URL
"finger://nasanews@space.mit.edu", not "finger://space.mit.edu/nasanews".

To avoid this confusion, I figured it was best to let them just put in the
request they would expect. My format lets them put in the /W and the
multiple hosts in a very predictable fashion instead of having them
remember which part went where: it goes in just as it would look on the
command line, or on the request that is sent to the finger server.

--Paul Hoffman
--Proper Publishing