Language Embeddings.

In order to use the kernels, the LINDA primitives have to embedded in some programming language. We chose two languages, an imperative one and a functional one, in order to assess whether the relative suitabilities and expressiveness of the tuple-space extensions. This outcomes of this part of the project were:

1. embeddings of the tuple space model in ISETL (a set-based, imperative language), and Gofer (a non-strict, purely functional language).
2. the conclusion that the addition of tuple-space functionality to ISETL integrated well, and caused little conceptual change.
3. embedding `stateful' tuple-spaces into a `stateless' language caused less difficulty than expected, but meant that recourse has to be made to unusual techniques (continuation passing style) which give an imperative `flavour' to declarative languages.