Errata of "Properties of Realistic Feature Models Make Combinatorial Testing of Product Lines Feasible"

1. The formula in Section 5.1, F^t = M/(t * x), is inaccurate. It should have been (f choose t) * 2^t = M/(t * x). Consequently, instead of "= 20,000 features" on p. 649, it should have been "≈ 14,143 features".

2. The algorithm on p. 644 does not say that the invalid tuples should only be removed once, and that is on the given condition. The algorithm still works but is slower. The condition should have been included in the pseudo-code and was included in the implementation all along.

3. We wrote on p. 643: "Therefore, for the class of feature models intended to be configured by humans assisted by computers, which we think at least is a very large part of the realistic feature models, quick satisfiability is also a property." This is an under-statement given the argumentation preceding it. It should have been: "Therefore, for realistic feature models, quick satisfiability is a property."

4. The sentence on page 103 and 107 ending with "an NP-hard problem." should have ended "an NP-complete problem without a polynomial-time approximation scheme (PTAS)." On page 108, two occurances of NP-hard should have been "NP-complete".


Please contact Martin Johansen in case you have found an error not listed here and the errata will be updated.