Observations on Spreadsheet Languages, Intension and Dataflow

Spreadsheet languages for distributed computing are of great interest because they unite ease of use with the ability to express parallel computations. This paper discusses some issues that these languages raise, in the context of dataflow and intensional programming languages such as Lucid. First, proper design and implementation of spreadsheet languages (as we see it) places stress on both the eductive and data-driven dataflow models of computation. Second, spreadsheets feature both relative and what we have called implicit addressing, which are both forms of intension; we compare these to the intensionality found in Lucid. Third, since spreadsheets are often used by unsophisticated users, the abstract reasoning skills taken for granted by scientists cannot be assumed. This elevates syntax in importance. We therefore propose an extended multi-dimensional array notation which unites simplicity, intensionality, concreteness and historical usage.