The Architecture Of a Computer Processor

we analyse the intuitive concept of reflectivity which directs our definition of an abstract notion of sequential deterministic reflective al- gorithms in Sect. 4.3. As characteristic examples for the practice of reflective programming we specify the reflective behaviour of programs in three con- crete (logical, functional and imperative) languages: Prolog (Section 4.4), LISP (Section 4.5), and the reflective version RASP (Random Access Stored Program Machine) of the RAM (Section 4.6). In Sect. 4.6.1 we show an ex- ample of a reflective machine at work, namely a RASP simulation of indirect addressing. The simplicity and the expressive power of our generic concept of reflective machines are also highlighted by using reflectivity to define a com- putationally universal abstract machine (Sect. 9.1.2.1, pg. 228). In Sect. 4.7 we compare the reflectivity features of the mentioned languages with our def- inition of reflective ASMs and relate them with the closely related concept of reification of programming features, a different reflectivity style offered for example by Java.

The Architecture

When operating on structured objects one might wish to permit the inde- pendent specification of two or more ‘local’ updates in parts of the structure that are then combined (‘aggregated’) into one ‘global’ update of the struc- ture, possibly also mediating conflicts that could occur among the given local updates. This consideration applies in full generality to multiple so-called partial updates for any structured objects, like sets, functions, graphs, trees, programs, and also to multiple assignments to shared locations in concurrent runs, e.g. a counter where updates by different agents may be intended to be applied cumulatively