7 Jun
2013
7 Jun
'13
6:13 p.m.
> While I would much rather have had the 5150 give
options of what to do
> when a parity error occurs, I can see that most people would just keep
> hitting "Ignore", and then later complaining how unreliable the machine
> is. Ignoring errors gets higher user-satisfaction ratings.
On Fri, 7 Jun
2013, Cory Smelosky wrote:
Or in my case, higher usability. Some of my projects
/require/ pressing
"ignore". I'm quite fond of unsupported configurations...
I would have preferred that the PARITY ERROR would give a CHOICE:
shutdown (the only choice that it had)
Which is usually not acceptable. I would rather have my word processing
document with knowledge that one or more bytes are corrupted, than LOSE
it. Can "modern" software handle a file that has a bad bit in it, or
will Weird and Excess just choke on it?
Save RAM to a FIXIT.DAT file and reboot.
drop into DEBUG (as demonstrated by NMI debuggers)
give a choice of Ignore?
Are you sure?
Are you REALLY sure?
Do you realize that there is KNOWN wrong bit(s) that you are choosing to
pretend are OK?
Do you promise not to blame US for your corrupted data if you choose to
continue?
AND, PARITY ERROR should always give the full exact address including the
current content/bit pattern.
--
Grumpy Ol' Fred cisin at xenosoft.com