Re: finger, again
Mon, 6 Mar 1995 12:45:23 -0500

Date: Mon, 6 Mar 1995 12:45:23 -0500
Message-Id: <>
To: (Paul Hoffman)
Subject: Re: finger, again

PH == Paul Hoffman <>

  PH> The current (and future) draft says: "Clients should not decode CR and
  PH> LF characters in a URL."

This could be interpreted to mean: Do not decode "%0[adAD]" in a request
string.  It might be better to say that any request with that string fails,
since otherwise implementors will have to put together a special decoding
routine for just this scheme.

  JRhine> Speaking of end-of-lines, does your finger proposal specify how
  JRhine> the <request> is terminated?

  PH> No, because...

  JRhine> Does the finger protocol spec?

  PH> Yes, it does: a CRLF.

I haven't read the finger draft in a while.  Does it say, "The sending of
the request to the host shall be peformed as per the finger protocol spec"
or does it say, "The client shall send the request string to the host, port
79, in 7-bit ASCII" (basically duplicating the finger spec information).  If
implementors require the finger protocol spec to properly implement the
finger scheme spec, that must be made very clear in the finger scheme draft.

-- | Harvey Mudd College |

"Remember, only users lose drugs." -- from Richard Stueven