[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Issue: FUNCTION-DEFINITION (Version 1)



My notes from Fairfax meeting...

Cleanup meeting:

 VAXLISP and Lucid have this function, but call it SOURCE-CODE.
 KMP agreed to name change.

X3J13 meeting:

 Beckerle: `Let 'em write hash tables.'

 Skona: This function would be very useful in teaching situations
	where people don't yet know about hash tables.

 KMP: Program modularity does not always permit the hash table
      solution. Also, the hash table solution has GC impact unless
      we have weak hash tables.

 Masinter: This would be an undue burden compiled-only implementations.

 KMP: Cloe is the only compiled-only implementation I can think of
      and doesn't consider it to be an undue burden as long as
      the info can be discarded by compile-file.

 van Roggen: VAXLISP now has several different kinds of evaluators,
	     and this might impact some of them. Wants to study issue
	     further.

 KMP thinks that some people seemed to have the mistaken belief
 that one of these proposals required source code info to files
 because they were worried about gratuitous bloating of dumped
 applications. Since both proposals only involve compile-to-core,
 they weren't nearly as wasteful of space as some seemed to believe.
 This needs to be made more apparent in the writeup.