Repository navigation

separation of floating point / fixed point numbers

currently uses double for all numbers, regardless if they're float or not
might increase performance
will increase complexity

switch / case

same style as in javascript / c, et al.

dynamic parser / extend and/or modify parser at runtime

requires the parser to be present even during execution, incurs overhead
a very primitive variation already possible with eval()
adding new keywords, symbols, et cetera, means dynamic table, rather than static
probably not very realistic

inline assembler / runtime registry execution

requires a parserfor a kind of assembly language corresponding to the register opcode shortcodes
could be used to alter code interactively, perhaps even self-modifying code
syntax maybe: ASM{ ... }
bit of a stretch, really

decorators / attributes:

[threads(4)]
function iterate(stuff)
{
    ....
}

class Blargh
{
...
}

[]