摘要
Software trustworthiness measurement can provide evidence for increasing the trustworthiness of the design and implementation of software. We once studied the application of axiomatic approaches in software trustworthiness measurement, presented seven desirable properties of software trustworthiness measure based on attributes, and established three measures. However, none of the three measures can fully satisfy these seven properties. In this paper, we propose an improved software trustworthiness measure based on a partition which satisfies all of the seven properties. Meanwhile, we conduct an empirical validation of this measure by a real case. Comparative study shows that this measure is better than the three measures we once built.