Re: Data Corruption / 400

Thanks for all the great information, that confirmed what I suspected.
I heard from several people off list, and from both the M_L and Evansville
lists.

5 hits of the kind I was asking about, reported in an aggregate of 240 years
of experience in the posts I've seen so far.

Often when business data gets corrupted we really do not know what caused it
& focus on fixing it.  Several people knew precisely what had caused data on
a 400 to become corrupted, such as a lightning strike, power outage with no
UPS in which it is remarkable that they recovered all the lost business data,
and several people reported hits where they did not nail down the cause.

When I mentioned "software bugs" I was not referring to who made the mistake
- programmer at IBM customer site, consultant, or original vendor of the
software (I have seen all 3), but I had not considered possibility of a bug
in an OS release, prior to it being found & fixed.  I must say that IBM fixes
bugs in releases extremely fast compared to everyone else finding & fixing
their bugs.

Yes, I have seen cross-linked files, but that was application management
error.

I am familiar with runaway conditions testing a program that turned out to
have a loop, and mountain of error messages from a flaky connection where
perhaps no one monitoring QSYSMSG or QSYSOPR.

In the non-standards world of confusion about what is proprietary & what is
open, the OS can get blamed for what might be a conflict with 3rd party
software mods to the OS reality, that I think the OS should not permit.  This
is something that seems much more prevalent in the Microsoft world than in
computers from companies that came from mainframe & mini & midrange, like IBM
& the 7 dwarves.

However it was interesting to hear from people with first hand experience that
Linux on PCs is very stable compared to the standard OS of PC & Mac.
One contact recently had some Oracle training in which learning how to
recover from corrupted data bases is part of that job.

In 1984 I was working at a place that had put something called RPL (I think)
on an IBM Series/1 which was like MAPICS in that everything was accessed by
address chains & if you had a disruption, the chains could get broken, and
you could not reconstruct your data.  When there was a weather front, it was
like pulling the plug on the remote users who were linked by a non-standard
(wire your own) ma bell connection.  That was the employer that went
bankrupt.  They knew enough to keep the mission critical stuff on S/34.

MacWheel99@aol.com (Alister Wm Macintyre) (Al Mac)