Software components need a uniform approach for rating their quality. The n
eed for this stems from a licensee's inaccessibility to the source code as
well as other information concerning how thoroughly the component was valid
ated. This paper proposes a 'test quality rating' (TQR) metric that will ac
t as a component's dependability 'score'. We envision a process whereby a s
oftware publisher submits a component to an independent certification organ
ization that would then calculate TQR for that component. It is preferable
for component dependability validation to be performed by an independent or
ganization. This is because even an honest dependability overestimation err
or on the part of the publisher could be grounds for severe legal penalties
. This score would be displayed on any marketing materials or contracts whi
ch license that component. In our paper we provides results from applying t
he metric to a commercial financial application written in Java in order to
demonstrate the effectiveness of the metric. (C) 2000 Elsevier Science Inc
. All rights reserved.