Received: from SOUTH-STATION-ANNEX.MIT.EDU by po7.MIT.EDU (5.61/4.7) id AA08014; Sat, 6 Jan 96 07:47:43 EST
Received: from anavrin.cs.titech.ac.jp by MIT.EDU with SMTP
	id AA19657; Sat, 6 Jan 96 07:46:42 EST
Received: by anavrin.cs.titech.ac.jp (8.6.12+2.4W/3.2W-titmx-2.0g); Sat, 6 Jan 1996 21:46:32 +0900
Message-Id: <199601061246.VAA01291@anavrin.cs.titech.ac.jp>
To: sipb@MIT.EDU
Subject: WWW HTTP question
Date: Sat, 06 Jan 1996 21:46:31 +0900
From: Monty <xiphmont@cs.titech.ac.jp>


Hi folks, I'm asking this over mail because I'm nightshifted and likely to be
online when no one at MIT is around...

I'm seeing a new weird kind of search string submission, which of course, the
Web server and Forum can't parse.  Forum requests are encoded into URLs as
such:

http://deskfish.cs.titech.ac.jp:8001/web-nntp/server?request

Standard search string stuff, where the request is something like:

action=fetch&spool=gardening&aptr=31&E

These request URLs are embedded into the documents that are served back.
Nothing alarming there.

However, there's a web browser out there that is sending back requests like
this:

fetch\?gardening\?31

It's consistently parsing the embedded URL into and abbreviated form,
eliminating the field name and using '\?' for the delimiters.  WTF?  Does
anyone know anything about it?  Seen it before?  I've been pouring over HTTP
specs and can't find anything like it... 

Monty
