Peter Bartholomew
Well-Known Member
@chirayu
True, but my reason for opening up the discussion was that not all errors are at the detailed implementation level. In fact the most damaging errors can be at the planning stage which draws on both domain knowledge and Excel solution logic.
Not that I would fault anything you have suggested.
One further check that I would propose is to count the rows of input data and then test whether every downstream calculation is adequately sized to accommodate it in its entirety. Much of the testing I have seen has the content of the data as its focus rather than the structure (if any) of the proposed solution.
Peter
True, but my reason for opening up the discussion was that not all errors are at the detailed implementation level. In fact the most damaging errors can be at the planning stage which draws on both domain knowledge and Excel solution logic.
Not that I would fault anything you have suggested.
One further check that I would propose is to count the rows of input data and then test whether every downstream calculation is adequately sized to accommodate it in its entirety. Much of the testing I have seen has the content of the data as its focus rather than the structure (if any) of the proposed solution.
Peter