Re: uh oh -- halp!

Bill Janssen <janssen@parc.xerox.com>
Message-id: <kgXfMeQB0KGWA2LO41@holmes.parc.xerox.com>
Date: 	Wed, 8 Sep 1993 21:50:50 -0700
Sender: Bill Janssen <janssen@parc.xerox.com>
From: Bill Janssen <janssen@parc.xerox.com>
To: ebina@ncsa.uiuc.edu, "William M. Perry" <wmperry@mango.ucs.indiana.edu>
Subject: Re: uh oh -- halp!
Cc: www-talk@nxoc01.cern.ch
In-reply-to: <29140.747522693@mango.ucs.indiana.edu>
References: <29140.747522693@mango.ucs.indiana.edu>
Status: RO
Excerpts from ext.WorldWideWeb: 8-Sep-93 Re: uh oh -- halp! "William M.
Perry"@mango (1336)

> When you make the connection to the server, it
> automatically spits it out as a 'hello' message.  Just have the
> browsers wait for one line of output, then parse it to see if it
> accepts http/1.0 requests, then send the appropriate request.  It
> might mean a slight slowdown for sites halfway around the globe, but
> then the connection won't be blazingly fast anyway. :)

>    I had always wondered why this wasn't in the HTTP spec, since most
> of the other protocols seem to use it (SMTP, NNTP, POP, etc).

One reason might be that this kind of wart makes it much harder to port
the HTTP protocol to a non-connection-based (say, message-based RPC)
transport layer.

Bill