Subject: Re: Support for earlier Personal Computer Adapters
To: Gregg C Levine <hansolofalcon@worldnet.att.net>
From: Greg A. Woods <woods@weird.com>
List: tech-net
Date: 08/12/2000 00:51:18
[ On Friday, August 11, 2000 at 23:00:13 (-0400), Gregg C Levine wrote: ]
> Subject: Support for earlier Personal Computer Adapters
>
> Are there plans to support such adapters, as the IBM 3278/79 adapter? It was
> built by them during the late 80s, and up to the middle 90s. Typically  the
> sotware would be written by them for connection to a S/3X0 mainframe
> computer, through a communications controller, as the IBM 3174. Sorry for
> the history, but I am involved in a project that uses them, and the customer
> had heard of NetBSD during his search for operating systems through the
> Internent.

In days *WAY* gone by (at least in my frame of reference, say about
1986-1989) I wrote some such "drivers", first in MS-DOS and then as a
DLL in M$-Win-v2.1, to use with such cards (eg. IRMA).  They're usually
pretty simple things to program and a unix driver shouldn't be all that
hard to write.  Some of the later cards had quite a few smarts onboard
and IIRC they had a dual-ported memory mapped region representing the
screen real-estate and maybe one or two I/O ports with which to control
local options and to trigger the various command functions.  I think one
variant I worked with even had an IRQ that would trigger when the new
screen was "painted" and ready to read.

We then went on to build client/server systems that used highly
customised screen formats to make the best possible use of screen
real-estate to transfer data to and from the mainframe.

I did do a very minor bit of experimentation with one of DCA's(?) latest
cards on an Interactive Systems i386 unix box (SysVr3.2, IIRC).  Our
plan was to centralise the mainframe interconnect on the unix server so
that we didn't have to deploy 3270 cards and cables everywhere (since
there was already a LAN in place too!).  I think there was already a
driver though and my task was going to be to write a communications
server process that would multiplex the client requests efficiently.  I
ended up taking another job though and I think the project was
cancelled.

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods@acm.org>      <robohack!woods>
Planix, Inc. <woods@planix.com>; Secrets of the Weird <woods@weird.com>