Persistent Programming

The effect of eliminating a separate file system is not new.  The aim has been pursued by Prof. Malcom Atkinson and Prof. Ron Morrison since at least 1983, with the publication of the paper "An Approach to Persistent Programming" [11]. This and their subsequent papers describe the advantages of eliminating the idea of conventional file systems from programming languages and describe how this can be implemented on conventional hardware.

The fundamental difference between their approach and the approach adopted in Timor is that they implement persistence in the compiler, without the aid of unconventional hardware and therefore with considerable difficulty, whereas Timor assumes the existence of unconventional hardware, such as that of SPEEDOS, which enormously simplifies the implementation. Appropriate hardware was first developed by Keedy and his team in the MONADS Project at Monash University in the late 1970s and early 1980s (see the Monads website https://www.monads-security.org/) but this not generally available. However, Keedy has recently written a paper which describes how RISC computers can easily be adapted (at the hardware level) [15] in such a way that a persistent SPEEDOS operating system and other persistent systems could be built, if RISC computer manufacturers make this modification.