Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!decwrl!ucbvax!NMS.HLS.COM!salzman
From: salzman@NMS.HLS.COM (Mike Salzman)
Newsgroups: comp.protocols.tcp-ip.ibmpc
Subject: Re:  Who has INT14 specs for network comms?
Message-ID: <9009260824.AA06525@nms.>
Date: 26 Sep 90 08:24:18 GMT
References: <9009292304.AA27382@alw.nih.gov>
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: Hughes Lan Systems, Mt View Ca
Lines: 25

As long as extensions to the basic spec are brought up.....

Roger Fajman is correct in stating that the original specs were
documented in the AT Tech Reference.  Aside from Bridge, others have
added extensions to the interface because it was so brain dead.

A common addition, first instituted by Walker, Richer and Quinn
provided a function call to initiate a BREAK at the port.  UB 
developed a completely different interface, INT 6B, for handling
blocks of data more efficiently across a network connection.

The problem (dillema) with interfaces is that a different 
code module is required in each case, since the mechanisms
are so different.  However, the original notion that INT14 
would allow a package to operate in a real PC, out the serial
port, or over a network connection, is not credible.  As any
one who has tried it knows, even an AT can barely keep up with
a 4800 baud connection on the port, if they use the BIOS
interrupt.  Most software interfaces directly to the 8530
chip, and bypasses BIOS.

-- 
-------------------- salzman@hls.com ----------------------
Michael M. Salzman  Voice (415) 966-7479  Fax (415)960-3738	
Hughes Lan Systems  1225 Charleston Road   Mt View Ca 94043