## Tuesday, January 20, 2015

### The upper bound on the number of test cases in the new regression test is 2,359,296, but it'll probably, hopefully, be less than that.

We're coming up with a new regression test at work that kind of, maybe,
combines both the original “Project:
Wolowizard” and the new “Project:
Sippy Cup” and my manager, S, came up with 11 variables to test; A with
values A_{1}, A_{2} ‥ A_{8}; B with values
B_{1}, B_{2}, B_{3}, B_{4}; C with values
etc. And multiplying out all the possible values of all the variables, I
end up with a potential 2,359,296 possible test cases.

Now, some of the cases aren't possible. For instance, values
D_{1}, D_{2}, D_{4}and D_{6} can
*only* be tested with value H_{1} (or rather, with those
values of D, H_{1} is the only value H can have), while
D_{3}, D_{5}, D_{7} and D_{8} can only be
tested with values H_{2}, H_{3} and H_{4}. So, just
looking at D and H, it's … um … let's see … there are two parties
involved, both with D and H values, and does it make sense for
P_{1}D_{5} to call P_{2}D_{5}? Um …
well, in this case, H doesn't really apply to P_{1} so out of a
possible 32 combinations, only … um … 24 really apply.

I think.

Yeah, it's going to be *real fun* trying to puzzle out all the
cases.