Recent developments have shown the feasibility of implementing sophisticated simulation language facilities on even a modest minicomputer configuration. Unlike conventional simulation languages which, in the main, operate in a non-interactive batch mode, minicomputers generally allow the user to simulate in a highly interactive manner. This paper discusses three strategies used by the authors to implement different versions of an interactive simulation language on a minicomputer, and describes the means of achieving highly interactive simulation features. There are two fundamentally different approaches to implementing a simulation language. The first, a translator approach, makes use of existing software, for example editors, high-level language compilers, job control language features, and the language is built using these software foundation blocks. The second approach is to construct the facility independently of existing software and to provide a package which is quite independent of the computer's operating system. This second approach can be further subdivided into two branches, the compiler approach, in which the object code produced by a simulation language compilation is the machine code of the particular computer, and secondly the interpreter approach, where the object code is machine independent and is subsequently interpreted by code expressed in a high-level language. The various implementation strategies are described and the advantages and disadvantages of each approach are discussed in relation to three implementations of a simulation language for PDP8 computers.