Phil wrote:
> then any Y2K problem uncovered could have ever been. The no. 1 experience
> I would take from this is: Take complete backup of disks before starting
> tests. Do *complete* restore from backup when finished. Test the backup
> set before starting the tests. The hours and money this obvious step would
> have saved is amazing.
Thanks Phil, sounds like good advice!
cheers,
peter
I would echo that.
Another key point of Phil's was "... generally the Y2K testing has been more
massively disruptive and costly then any Y2K problem uncovered could have
ever been."
If you are going to do date warp testing (as it is called) then be very
clear about what specific bugs you intend this testing to uncover. They must
be bugs that you couldn't have discovered in normal testing - otherwise much
cheaper, easier, less disruptive to find them under current date conditions.
Clock interactions with processing and data are the key (seems obvious - but
based on experience ...), boundary conditions, scheduling and interfaces
also.
Remember, testing doesn't prove the absence of bugs only their presence.
---------------------------------------------------------------------------
New Zealand Delphi Users group - Offtopic List - [EMAIL PROTECTED]
Website: http://www.delphi.org.nz
