If you look at the architecture of a computer, and the history of computer development, it is a designed constuct that operates on logical principles, has a clear data representation strategy, and was developed step by step into methods for combining low level physical processes in such a way that they can be described and manipulated in high level abstract symbolic terms. Computational systems analysis is a high level data and logical abstraction analysis. It can produce data flows and algorithms for manipulating them. Programming uses software that is based on logical manipulation of lower level computational abstractions. Those programs use data that is stored in logical and separable ways in the computer, including magnetized regions in conventional hard drives. The lower level processing software uses assembly language, which itself is a collection of routines that translates into machine code. Machine code in modern digital computers is binary, 0 or 1. This is what is encoded in various architectures. The brain is very different. Things are not stored as 0 or 1. That storage can be modified, sure, but so can any physical structure. Modifiability is not enough to proclaim it software, nor at the software level. Yet it might be argued that the brain can modify its own structure, and that therefore this is equivalent to that structure being software. This would be a mistake however. Much of what the brain does is instantiated and mediated by local physical processes. Its not run according to any symbolic or rules architecture. There is no software processing of such structure, neither by symbolic processes nor logical algorithms. Abstracting things to a software description is just a model. Again it might be argued that we use symbols, and we think in terms of symbols, which demonstrates a software processing capability. Again, this is a mistake. Modern neuroscience is showing most of what we do, how we think, is non-symbolic. The very "symbols" we process are not precisely developed definitions/data, but flexible distributed processes, most of which is nonconscious. Because of higher brain function we can learn and manipulate rules. In a very limited sense this can be said to be what a computer does, but the nature of how we do it is so very different. We are still learning about that. We are not logical or rational beings, we do not think symbolically. We are creatures who can learn to use logic or rationality or symbols, just as we can learn to speak English, or Dutch, or do higher mathematics. There is an abstract level that can describe this in information theoretic terms, but such terms miss the complexity of what is going on. It is when someone starts thinking of such abstractions as real processes in the brain that mistakes are made. What can be changed, and how it can be changed, that could be described as learning or adaptation, is itself defined by brain architecture. We are only just beginning to understand this. I regard the view that the brain is a computer as a major impediment to progress. These are fundamental problems relating to abstraction and the nature of models. The most important statement in systems theory is this: the map is not the territory.