the study of programming languages

ryan stansifer

pat scanned it. thanks pat.

Simplicity enters in four guises: uniformity (rules are few and simple), generality (a small number of general functions provide as special cases a host of more specialized functions, orthogonality), familiarity (familiar symbols and usages are adopted whenever possible), and brevity (economy of expression is sought). p. 20

thoughts:

the computer understands only bits, or understands only self-referential statements
the point of shannon/turing is that it doesn’t need to understand anything
in the end, it only understands on or off.
it makes distinctions between entities and the name of entities (called the use-mention problem)
the association of a name to an attribute is called binding
- one thing that the computer/compiler knows but not the programmer is the location of everything

computers understand structure and humans understand names one example of this separation is when you check for type equivalence: either you check for structural equivalence, or your check for name equivalence. because it’s easier to reason about, modern programming languages choose name equivalence. (p.118)

typoologies of programming languages:

imperative programming
- moving data around in storage locations, dominating until the 80s
- it’s all about assignments -> place value into storage, most statements end up in that

pointer semantics:

identifier can point to a location (environment)
a location can point to a value (state)

a language without contingency, in which the context does not affect the meaning of expressions is said to be referentially transparent aka the meaning of a sentence must remain unchanged when a part of a sentence is replaced by an expression having the same meaning (frege, über sinn und bedeutung)