The Philosophy of Computer Science
If you need philosophy, please break the glass with an axe.
Excerpt from here ( http://plato.stanford.edu/entries/computer-science/ ):
-------------------------------
2.1 The Dual Nature of Programs
Many authors (Moor 1978; Rapaport 2005b; Colburn 2004) discuss the so-called dual nature of programs. On the face of it, a program appears to have both a textual and a mechanical or process-like guise. As text, a program can be edited. But its manifestation on a machine-readable disk seems to have quite different properties. In particular, it can be executed on a physical machine. So according to the principle of the indiscernibility of identicals (§3.3), the two guises cannot be the same entity. Of course, anyone persuaded by this duality is under an obligation to say something about the relationship between these two apparent forms of existence.
One immediate suggestion is that one manifestation of a program is an implementation of the other i.e., the physical manifestation is an implementation of the textual one. However, even within the confines of computer science, it is not immediately clear that the word implementation refers to just one notion. Often it is used to refer to the result of a compilation process where a program in a high-level language (the source code) is transformed into machine language (the object code). But equally often it is used to refer to the process where the source code is somehow directly realized in hardware (e.g. a concrete implementation in semiconductors). And presumably, this is the relevant notion. But without a more detailed philosophical analysis of the notion of implementation (§3.2) itself (Rapaport 2005b), it is unclear how this advances the discussion; we seem only to have named the relationship between the two apparent forms of existence. In a similar vein, others have described the relationship between the program-text and the program-process as being similar to that between a plan and its manifestation as a series of physical actions. But this does not seem to be quite analogous to the program-process pairing: we are not tempted to refer to the plan and the physical process as being different manifestations of the same thing. For example, are we tempted to think of a plan to go for a walk and the actual walk as different facets of the same thing?
Perhaps matters are best described by saying that programs, as textual objects, cause mechanical processes? The idea seems to be that somehow the textual object physically causes the mechanical process. But this would seem to demand some rather careful analysis of the nature of such a causal relation. Colburn (2004) denies that the symbolic text has the causal effect; it is its physical manifestation (the thing on the disk) that has such an effect. Software is a concrete abstraction that has a medium of description (the text, the abstraction) and a medium of execution (e.g., a concrete implementation in semiconductors).
A slightly different perspective on these issues starts from the question of program identity. When are two programs taken to be the same? Such issues arise for example in attempts to determine the legal identity of a piece of software. If we identify a program with its textual manifestation then the identity of a program is sensitive to changes in its appearance (e.g. changing the font). Evidently, it is not the text alone that provides us with any philosophically interesting notion of program identity. Rather, to reach an informed criterion of identity we need to take more account of semantics and implementation. We shall return to this subject in §3 and §6.
