Path: utzoo!utgpu!utstat!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!lll-winken!scooter!neoucom!wtm
From: wtm@neoucom.UUCP (Bill Mayhew)
Newsgroups: comp.sys.amiga
Subject: Re: multitasking on wallstreet
Summary: IBM clone RAM parity chip is almost worthless
Keywords: IBM PC RAM parity
Message-ID: <1526@neoucom.UUCP>
Date: 9 Mar 89 16:15:07 GMT
References: <1295@csd4.milw.wisc.edu> <503@morgoth.UUCP> <32244@auc.UUCP> <326@bnr-fos.UUCP>
Organization: Northeastern Ohio Universities College of Medicine
Lines: 29


Unfortunately, once you get an interrupt from a parity error on a
PC clone, the architecture of the machine itself prevents your
program from taking much of an action other than cold rebooting.
To this extent parity checking is not extraordinarily helpful.

Unfortunately at the time the original IBM PC-1 was introduced,
there weren't any good inexpensive EEC controller components
available.  Now it would be reasonable to make a memory with EEC so
that errors could be detected and most corrected.  Nobody has
bothered on a PC clone yet as far as I know.  Parity chips as used
on PC clones are a good way for memory companies to increase their
sales volume by 12.5%.

I don't find the lack of a a parity bit offensive.  Any decent
package that is transferring critical data darn well better have
some sort of embedded hash digit checking at the software level.

Where you get into trouble is when a character is sitting in video
ram and a bit glitches causing you to see the wrong thing on the
screen.  When did you ever see parity bits on video boards?  In the
IBM world, you could get burned (when not in bit-mapped mode).  On
the Amiga we win becasue the characters are rendered onto the
bitmap; a one-bit glitch is not going to make a 8 -> 9 as it could
on a character display.  ... and you could tell something was up
as the character would look funny.


--Bill