Key points are not available for this paper at this time.
In programming languages which permit both assignment and procedures, distinct identifiers can represent data structures which share storage or procedures with interfering side effects. In addition to being a direct source of programming errors, this phenomenon, which we call interference can impact type structure and parallelism. We show how to eliminate these difficulties by imposing syntactic restrictions, without prohibiting the kind of constructive interference which occurs with higher-order procedures or SIMULA classes. The basic idea is to prohibit interference between identifiers, but to permit interference among components of collections named by single identifiers.
John Reynolds (Sun,) studied this question.