Macro testing
When you develop programs, you have to test, test, test. After this past month, I see at least 2 areas where you have to do some serious testing:
1) Test different sites/units/businesses that will be using your program. Each site or unit may have different ways of entering data or have different terminology. Even the download layouts could vary. If you are able to test all of the sites, then do so. If you have thousands of them, then that’s not achievable but if you have 50, you may want to test them.
2) Test across time, especially if your program has a time element. In particular, I would pay attention to monthly switchover or yearly switchover. You might come across unexpected results. Currently, the OneSite download will show me Jan-15 and I think, January 2015, but it’s really saying 1/15/2014 if you try to apply formulas. Ouch! I had to do some major rewrites when I recently encountered this problem, only because I am starting to draw data for 2015.
Aside from calculation errors, those are the two things that will most likely trip you up and they will be hidden unless you do a lot of testing.
Of course, in about a week or so, I’m going to find out a lot more trip ups because that will be when I start user testing.
You must be logged in to post a comment.