Setting up a VAXstation - test-drb@ccmp.vtda.org

8 Oct 2007

...

Subject: Re: Setting up a VAXstation
   From: shoppa_classiccmp at trailing-edge.com (Tim Shoppa)
   Date: Mon, 08 Oct 2007 08:14:33 -0400
     To: cctech at classiccmp.org, cctalk at classiccmp.org

Tony Duell said:
  I cna understnad why people are interested only
in old software, not
 hardware, and want to run it under emulation on a modern machine

 My puzzlement is with people who want to run the old hardware (not have
 to run the old hardware becuase it is part of some machine tool or
 something) but don't want to understand what's going on inside. What more
 do you get over running the software under emulation? 
In fact, availability of hardware is a huge factor in succesfully
making an emulator for a machine. All but the simplest processors
are complicated enough that there are little corner cases all over
the place where none of the processor/architecture documentation tells
you what is going to happen.

And outside the central processors, all peripherals but the very simplest
are filled with complicated and undocumented behavior.

Schematics could answer many of these questions, but in real life
they end up guiding the search for the answer to the question rather
than being the actual defining source for the answer.

So in general emulator users and especially developers completely
grok the need to have hardware working. Sometimes I believe that
today's emulator developers know much more about the architectures
than the original architects did :-). (In a couple cases, they are
the orignal architect!)

Availability of software is also important for making a reliable emulator.
You could spend years reading the books to write an emulator, but you
don't trust anything you've read or done until you've booted the simplest
OS.

Tim. 
The best example of this that comes to mind is the Apollo Guidence 
Computer (AGC).  There was one hardy and persistant soul that not 
only researched it, he built a sim and tracked down samples of 
software to validate the sim and the later hardware.  One great
issues was lack of documentation, apparently much was lost/destroyed
when that chapter of the space program ended around 30 years ago.
The few intact copies of the AGC (Apollo command modules) likely 
haven't seen power in at least that long if even complete.  I doubt
any of the CM holders could be convinced to power it up assuming 
they were even preserved sufficiently to safely do so.  So for 
those interested in machines that are obscure, unusual or very 
rare even generating a sim has to be a huge challenge only equaled
by task of gatherering the needed data to base it on.  It is reverse
engineering on a very deep level for a faithful sim.

Allison