|
PureBytes Links
Trading Reference Links
|
Peter wrote:
> It is not a bug. It is simply the way all software reads dates.
Sorry Peter - it IS a bug. Note that there are two types of bugs:
implementation bugs (where the software fails to do what the designer
intended), and design bugs (where the designer's algorithm or intended
solution is defective). The Y2K problem is the latter type. Also, it
is NOT true that "all software reads dates" this way. All programs I
have written store dates as pseudo-Julian dates. So although early
programs did read input years as two digits, it was VERY simple to
modify the read statements to read four digits - NO change was required
to databases nor to other programs reading those databases.
Secondly, I feel that your priorities are wrong in stating that "Most
unwise when making major changes to do more than one thing at a time.
Better to do all the changes for TS5.0 and then go back and make all the
changes for the Y2K fix."
This is an example of "featuritis", the tendency to place much greater
emphasis upon adding new bells and whistles rather than getting what
already exists to work reliably and robustly. Unfortunately, featuritis
is a disease which infects most software vendors because unsophisticated
buyers are dazzled by all the advertising claims for the sexy new
"features".
Regards,
Carroll S.
|