On Sun, Feb 10, 2008 at 01:32:49AM -0800, Andrew Lentvorski wrote: > Agreed, *broken* Scheme/Lisp implementation are a dime a dozen. How > does quoting work? How about dotted pairs? How about quasiquoting? > Macros? Tail recursion via continuations? Lexical closures? *Correct* > implementations that are a reasonable number of lines are rare.
Incomplete definitely but not necessarily broken. We need conformance testing suite to shore up all the Scheme implementation floating around. That would be useful. > Similar things hold for continuations or macros--"Well, those > are too hard so were just going to use those from the implementation > Lisp". ARGGGGHHHH! > > "Best", in this case, means a small number of lines combined with a > straightforward implementation that covers all of the essential features > --closures, continuations, tail recursion, macros (I'll allow simple > macros rather than hygienic). Maybe it can't be done in under 1000 lines with all the plumbing because those feature are nontrivial. Why not just suck it up and go look at the source to some GNU code like GNU Common Lisp? > This is my experience with functional programmers, in general. > Functional programmers implement the easy 80% and stop at the first sign > of pain while the C folks roll up their sleeves and painfully dig their > way to 90%+. http://www.gnu.org/software/gcl/ <- Go. Read. Quick yer belly-aching. :) Chris -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
