Re: Shunned ports [was: Predraft of a new URL scheme: mailmsg ]

Marc VanHeyningen (mvanheyn@cs.indiana.edu)
Sat, 07 Jan 1995 23:01:33 -0500


From: Marc VanHeyningen <mvanheyn@cs.indiana.edu>
To: Ned Freed <NED@innosoft.com>
Cc: "Daniel W. Connolly" <connolly@hal.com>,
Subject: Re: Shunned ports [was: Predraft of a new URL scheme: mailmsg ] 
In-Reply-To: Your message of "Sat, 07 Jan 1995 15:35:36 MST."
             <01HLKJ80AFZ68ZDW0P@INNOSOFT.COM> 
Date: Sat, 07 Jan 1995 23:01:33 -0500
Message-Id: <394.789537693@moose.cs.indiana.edu>

Thus wrote: Ned Freed
>> Couldn't this be generalized a little bit? Can't we assume, for
>> example, that the HTTP service is _always_ either on port 80, or on
>> some port >1024?
>
>> Similarly, gopher is _always_ either on port 79 or on some port >1024.

Methinks you mean port 70.

>This sounds like a good idea to me -- simple, effective, and covers future por
>t
>allocations cleanly.

I call this the list of Safe, Approved Ports, or the SAP list, (as
opposed to the list of Dangerous, Insecure Ports, the DIP list.)  We
can now have a debate of the dips versus the saps. :-)

>> This rules out the gopher: hack to access finger info. I won't loose
>> any sleep over that. A finger: URL should probably be deployed, along
>> with gateway interim solutions.
>
>Agreed.

Unfortunately, there exist FAQs and guides telling people to use
gopher URLs for finger access, whois access, and even HTTP access.
Client writers will ignore these guidelines rather than break these
existing URLs.

I agree that SAP is technically superior, but I think DIP is more
realistic.
--
Marc VanHeyningen  <URL:http://www.cs.indiana.edu/hyplan/mvanheyn.html>