These functions do not match XL 2000.

For example, Excel 2000 (Windows) gives

COUPNCD ("28-Nov-05", "29-May-09", 4) = 29-Nov-05

but Gnumeric gives 28-Feb-06.  Gnumeric with "TRUE" as the fifth
argument gives the (correct) 29-Nov-05 too.

Seeing as the buggy behaviour of the old Excel is fixed here, I
propose we drop Excel compatibility altogether.  Otherwise we're
chasing at least 3 targets, which is IMO pointless:

1) Excel 2000 (Windows)
2) Excel 97 (Windows - I assume this is where the big regression tests
   come from)
3) At least one version of Excel for the Mac.

And it appears they all have different bugs and fixes.  I propose we
just DTRT all the time.

Jody has proposed we just chase the latest Excel for Windows.



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]