Automated recognition of features from CAD models has been attempted for a wide range of application domains. In this paper we address the problem of representing and recognizing the complete class of features in alternative interpretations for a given design. We present a formalism for representing feature- based design alternatives and a methodology for recognizing a class of machinable features. Our approach handles a class of volumetric features that describe material removal volumes made by operations on the three-axis vertical machining centers including: drilling, pocket-, slot-, and face-miling, chamfering, filleting, and blended surfaces. Our approach recognizes intersecting features, and is complete over all features in our class, i.e. for any given part, the algorithm produces a set containing all features in our class that correspond to possible operations for machining that part. This property is of particular significance in applications where consideration of different manufacturing alternatives is crucial. In addition, we have shown that the algorithms are, in the worst-case, euqdratic in the number solid modeling operations. This approach employs a class of machinable features expressible as MRSEVs ( a STEP- based library of machining features). An implementation of these algorithms has been done using the ACISsolid modeler and the NIH C++ class library.