Prev: Fascinating interview by Richard Stallman at KTH on emacs historyand internals
Next: Emacs Time Line, Graphical Chart by Jamie Zawinski - Valuable Resource for Newbies - written: 8-Mar-1999, updated: 29-Oct-2007
From: Jacko on 18 Jul 2010 19:27 http://groups.google.co.uk/group/comp.lang.forth/browse_thread/thread/ef39e59023a8ae4c/9eb9f6b99139a5df#9eb9f6b99139a5df If (X . Y) has to be limited, as Y may on free assign limit it due to TLB and other things. Imagine if the link address did not need a virtual memory lookup, or could be loaded from a seperate memory bank. This can only be hartdwired for speed and power efficiency if ( . ) is removed. Removal is harder than adding. QED. It can be proved that garbage collection is faster and more localized if non-circularity is forced, and as circularity can be replaced by self reference then, not too much is lost for the big future gains. Such a constriction proof could have major impact in the future of processor design, and dogma of feature need my limit the performance for all, then some, then the few which use the old emulation interface. Constriction proofs and translation to 'acceptable' source may be the future, in a constraint for certainty processing model Why is (X (Y)) so bad? |