[<<][compsci][>>][..]Mon Jun 15 14:07:43 CEST 2009
The previous post advocates the explicitness of all references.
Instead of using this just for temporary file management, can't we use
it for _all_ file management? Can we view all files as temporary?
This is a shift in paradigm about how to think about data: instead of
looking at data as a dumb collection of bits, implicitly connected to
a program that uses it (an interpreter), you never disconnect it from
its use (semantics).
More generally, it sort of advocates the abolishment of the principle
of late binding. Goodbye Smalltalk? Actually, it seems to put late
binding into a clear perspective: it acknowledges data use cannot be
However, it seems to be exactly that which makes turing machines so
hard to understand. This is like making the leap from turing machines
to (static) boolean circuits as is done in complexity theory to make
the subject more managable.