pondering things

thoughts.markdown

• subexpressions juxtapose against what?

• locals (and variables in general) mandated by the interpreter, or an entirely libspace construct?

• is \n or ; handled by the interpreter, or implemented in libspace?

  • another way of putting it: are our AST-scopes a single line, using nesting for “subsequent statements,” á la S-EXPs/lisp?

• what is the meaning of the resumption value when given to a pristine execution?

• what sort of state do we have to store to juxtapose “previous value” against resumption value?

  • I don’t like storing state in executions, I didn’t think we’d need anything except a pointer into the AST.

• how to get multiple parameters to native routines before we can create multiple-element lists (affix(a_fork, descendant) takes two arguments, but we need affix() to build an arguments-list with two arguments, so we have to get around this)

  • currying-based native APIs? (affix(a_fork)(descendant))
    • difficult to implement for the interpreter
  • coroutine-based native APIs? (affix(a_fork); affix(descendant))

• nucleus is getting fairly solid, we should start writing tenative examples of Core-level code to sanity-check our nuclear designs

  • assumed-non-recursive implementation of default juxtaposition routine. need to worry about:
    • receiving parameters from Nucleus
    • no local variables at this level of abstraction
    • branch-and-stage handling (no calling at this level of abstraction)
    • no calling convention at this level of abstraction