For a technology founded on the automation of structured methods, CASE at present is singularly unstructured. There are now a myriad CASE tools available, all claiming to support a variety of methods, and all doing so in a variety of ways, good, bad and indifferent. It is therefore incumbent on the prospective user of CASE to evaluate the product (and all “upgrades” to it) extremely carefully.
The first step in this process is to discard the vendor's marketing claims, and to prepare your own very detailed evaluation procedure, based on your own development process and methods. CASE tools will not solve all of your problems: the best you can do is to establish what they will and will not do for you, and use them accordingly.
This paper illustrates an approach to evaluating tools that claim to support the Hatley/Pirbhai requirements specification method, showing the main criteria that should be investigated, and the required level of detail.