Software used to be written as strings of numbers, then acronyms. Now we use language-like expressions. A more important paradigm shift will be when 3D objects and behaviors - referents from reality, become the instruction set of software.

Current software has a very difficult task, it has to deal with complex objects in the real world, which it doesn't really have the machinery to cope with. it's a bit like watching little wooden toothpicks trying to manipulate huge boulders. They keep rolling the way they want, the tooth picks getting crushed underneath.

When you see the computer mouse close in on a button target, you see it as a complete animated 3D scene, the computer cannot see it at all, it has to logically test the mouse coordinate against each button coordinate until a match is found.

Computers use little bits to 'talk' about little functions, minds use millions of bits to 'talk' about those little bits. There is a gulf between the two. When a human thinks of a for-next loop, billions of parallel 'bits' are needed to handle the concepts as learned animated beliefs. But a computer needs only a tiny few. When a computer 'thinks' of a 3D object, all its 'effort' is in just holding the thing together. It's not generally seen as a computational unit in its own sense. - It needs to be.

Humans never use tiny bits to understand anything. There is no equivalent of ADD, SHIFT LEFT, COMPLIMENT etc. in the human mind. Every simulated object is made up of many thousands of parallel bits of data with many thousands of linked associations to other parallel sets of data.

Any subsequent thought processes are really meta computations using large data objects animated over time. Even to understand a simple XOR operation requires a 3D animation or at least a linked pattern sequence.

When we can model the procession of bits using animated 3d objects or pattern beliefs, we will be able to handle software development the same way we handle reality in the real world. These meta objects becoming the representational models of regular computer instruction code. They can then be properly simulated and used as the building blocks of robust computer generated code.

Common human language is much too ambiguous to control software processes. This is because it evolved primarily for the communication of emotional states, and then later, to include the description of objects and behaviors in reality.

Actual software code will be produced by translating the animated simulations of the computer code to actual binary code. The same way human language is primarily derived from interpreting animated scenes of reality. The animations will need to be 'emotionally' graded according to bit processing efficiency, code re-usability and speed etc.

Research Challenge:

Take a commercial 3D animation package like 3D Studio Max and develop a software interface layer that can leverage the power of the package to perform visual computing functions, such as align surfaces; align vertices; merge scale; merge form; break apart; join; equalize perspective; exaggerate motion vectors; clone behavior; fit to shape; enclose form; substitute behaviors etc.