> On Dec 6, 2016, at 11:12 AM, Pontus Pihlgren <[email protected]> wrote:
>
> On Mon, Dec 05, 2016 at 12:05:39PM -0500, Paul Koning wrote:
>>
>>> On Dec 4, 2016, at 10:12 PM, Bob Supnik <[email protected]> wrote:
>>>
>>> Josh Dersch at the Living Computer Museum found a snippet of real DECtape
>>> code (that runs on a real 11/40) which fails on the simulator.
>>> ...
>>> This leads me to think that there's a second principle to bear in mind when
>>> simulating older machines. The first is "economy of gates". In early
>>> systems, gates were precious, and the hardware tended to implement no more
>>> than the minimum functionality required. Error checks were a luxury and
>>> were often omitted.
>>
>> This makes me wonder about the fuzzy line between quirky features and
>> sort-of-bugs. The code snippet you mentioned sounds like it falls on the
>> side of the "quirky", and it sounds right for the simulator to implement
>> that. On the other hand, there's one I recently read about a machine in
>> which a subroutine call instruction would fail with the stack pointer equal
>> to -0, but when the stack pointer was +0 it would produce an address error
>> only for some of the addressing modes. "The schematics ... confirm this;
>> the reason is unknown" says the article. Implementing that sort of corner
>> case is obviously doable, but not necessarily all that useful.
>>
>
> Are you suggesting that simulators should fix "bugs" for someone using a
> simulator for comparison when restoring real hardware it could be very
> confusing.
>
> (not that such comparisons should be truated anyway)
No, I'm suggesting that simulators need not implement every known bug, if that
bug is unlikely to be exercised by real software, and if it's documented.
paul
_______________________________________________
Simh mailing list
[email protected]
http://mailman.trailing-edge.com/mailman/listinfo/simh
