Re: New Internet-Draft: finger URL

Reed Wade (wade@cs.utk.edu)
Wed, 15 Feb 1995 13:18:41 -0500


Message-Id: <9502151818.AA14209@honk.cs.utk.edu>
To: ietf-lists@proper.com (Paul Hoffman)
Cc: Reed Wade <wade@cs.utk.edu>, uri@bunyip.com
Subject: Re: New Internet-Draft: finger URL 
In-Reply-To: Your message of "Wed, 15 Feb 1995 09:00:49 MST."
             <v0211010bab67d0f2d735@[165.227.40.18]> 
Date: Wed, 15 Feb 1995 13:18:41 -0500
From: Reed Wade <wade@cs.utk.edu>


Michael A. Dolan (MD) and Paul Hoffman (PH) and Reed Wade (RW) write:


MD> Gee Wade, I thought I was helping by finding your old email..... :-)

Actually, it was a huge help, thank you very much.


MD>If the host is required, then why make yet another syntax 
MD>that must parsed ?

If the syntax were more complex I'd feel differently.
The ease of use on the users' end makes up for it, I think.


MD>Also, in the other scheme, how would one specify a non-standard port ?

You couldn't. But are there any instances of someone using another
port? This is certainly an arguable point, I don't have strong feelings
about it.


  >RW>There's one thing I'd like clarified--will the final @hostname
  >RW>part (from which the connect-to host is derived) be stripped
  >RW>before <blah> is sent to the remote host?
  >
PH>Unless I'm misunderstanding section 2.4 of RFC 1288, the finger client
PH>shouldn't strip anything. It looks to me that none of the "@hostn" are
PH>local to where the client sent the request from, and thus shouldn't be
PH>stripped.

The important question is whether I loose any functionality. Imagine
these 2 situations:

1. I type at my command line "finger wade@honk.cs.utk.edu"; the 
   finger command on my machine opens a tcp connection to the finger 
   server at honk.cs.utk.edu and sends "wade".

2. With finger:<wade@honk.cs.utk.edu> how can I send a "wade" to
   the finger server at honk?

   The finger server on honk may or may not treat the 2 requests
   ("wade", "wade@honk.cs.utk.edu") differently. We can't assume 
   they would do one thing or the other.


The rfc describes what goes over the wire, not the finger client
on any system. It's important to decouple the connection information
and the data part. I think that's the reason finger://<blah>/<blah
has the appeal it does.


-reed

-----
University of Tennessee, Knoxville            Dept of Computer Science
Netlib Development Group            'I was kidding,' says bomb suspect
wade@cs.utk.edu -- <URL:http://www.netlib.org/utk/people/ReedWade.html>