Re: Puzzle Solution

[Tom Droege <droege@fnal.gov>]

Chris and all,

Well, we still don't know if the bad ones (memory cards) are the good ones,
or the good ones are the bad ones or the bad ones are the bad ones, etc..
This is using a card that passes my random byte memory load and read back.
Actually I had to do two things to "fix" the problem.  One was to put a
dummy read at the start of the read out, and the second was to reverse the
byte order for putting together the word.  Both were done in the Basic code
that reads the memory and writes out a a file.  

It may well be that swapping in a "bad" card would fix the proble.  Some day
we will sort this out.  I don't (at the moment) much care what the problem
is as long as it is consistent.  It may be that the test path is really
different from the normal path.  We (Merle) did our best to make the
differences as small as possible. 

Still, note that Andrew had no more information than the rest of you
(possibly he was helped by knowing less than some of you so he was not so
confused) and solved the problem from first principals.  Note that I knew
that there could be a memory byte problem, yet it took me a week or so to
recognize it, working full time every day.  

I really like this problem.  It was so deliciously devious.

Tom Droege

At 10:51 AM 10/20/98 -0700, you wrote:
>Tom Droege wrote:
>> 
>> I would like to congratulate Andrew Bennett for correctly solving the
>> puzzle.  Below is his complete solution:
>> 
>> >
>> >It looks a lot like you are getting the wrong high order
>> >byte - either the one before or the one after.
>
>We still don't know _how_ it happened.
>
>You statement about a software fix threw me off.  I thought you
>meant STAMP software but I forgot you were using BASIC to read out
>the RAM buffer card.  Looks now like an "off by one" problem in
>the RAM buffer and then an incorrect byte swap.
>
>Do you know yet which stack of buffer cards contain the broken ones?
>Looks like a card swap would have fixed the problem also.
>
>-- 
>--Chris Albertson
>
>  chris@topdog.logicon.com                Voice:  626-351-0089  X127
>  Logicon RDA, Pasadena California          Fax:  626-351-0699
>