And.....

Support for negative numbers as they apply to dates prior to Dec 31 1899
should also extend to negative TIME calculations *without* having to change
to the 1904 date system. (which can cause other problems)

I'm thinking that this is rooted so deeply in the code and has been
propagated to so many other areas that it could never be fixed without a
"Y2K" effort. Read: money needed to fix it!

Biff

"Lex_Muga" wrote in message
...

Fred proposal could be one way of fixing the bug. I also support the
introduction of negative numbers before Dec 31, 1899.
--
Felix Muga II
Mathematics Department
Ateneo de Mania University
Philippines


"Fred Smith" wrote:

While it's understandable that MS wanted to have compatibility with
Lotus, they
could have preserved the compatibility and also fixed the bug.

To allow for an easy transition, MS wanted to ensure that, for example,
date
serial number 31,406 was Dec 25/85, as it was in Lotus. So they
programmed for
this.

Now the only problem is with the first 60 days of the 20th century. In
Lotus,
date serial number 1 is 1/1/1900 and date serial 60 is 2/29/1900, when
there is
no such date.

All MS had to do is say date serial 1 is 12/31/1899 and 60 is 2/28/1900.
Voila,
problem solved for all posterity. Yes, it would have caused a problem
with any
spreadsheet at the time that used the first 60 days of the century. But
the
number of affected spreadsheets at the time would have been
infinitesimal, and
not worth worrying about (as they could just stay with Lotus).

I still think MS should introduce this fix, and at the same time support
negative numbers for dates prior to Dec 31/1899.

--
Regards,
Fred


"
wrote in
message ...
"Dave Peterson" wrote:
since Lotus 123 thought that 1900 was a leap year,
that MS had a choice--do it correctly or do it to match
Lotus 123.
By matching Lotus 123's mistake, it would be easier
for users to migrate from 123 to Excel.
And since every company wants users and wants to
make that transition as easy as possible, the decision
was probably very easy to make.

As a developer of system software, I understand the
importance of compatibility. But assuming the explanation
is correct, there should be an option to correct the problem.
The default could remain bug-for-bug compatibility.