notes.txt

* Perhaps we could also implement a variant of lisp, where evaluation of an unknown symbol just stay the symbol itself (like in Wolfram Mathematica)

* bootstraping the compiler in the sense, that we can use rewritings to update the language implementation would be a cool idea. it would make it independent of C++ !

* as a note to the parallel flowboard project:
  lisp offers quasiquoting for generating code.
  however, we often not only want to produce code with generated components,
  but also "pattern match" to code with specific structure. it's simple
  to pattern match to code with variable sections to describe code transformations.
  however, expressing the same via functional access patterns and matching
  via == is cumbersome and not the best we can achieve.
  primary inspiration is scalas pattern matching and PROLOGs matching.

* profiling with: g++ -pg MiniLISP.cpp -o MiniLISP && echo "(...)" | ./MiniLISP | gprof MiniLISP gmon.out | head -n 50