Re: Cell execution dependence

*To*: mathgroup at smc.vnet.net*Subject*: [mg61982] Re: Cell execution dependence*From*: carlos at colorado.edu*Date*: Sun, 6 Nov 2005 19:46:26 -0500 (EST)*References*: <dkks4q$r91$1@smc.vnet.net>*Sender*: owner-wri-mathgroup at wolfram.com

Would this then be correct (in BNF) <whitespace-char> ::= <active-whitespace-char> | <inactive-whitespace-char> <active-whitespace-char> ::= (blank | tab | linefeed ) & activity-flag:ON <inactive-whitespace-char> ::= (blank | tab | linefeed ) & activity-flag:OFF If this is OK, the main interpreter burden for this classification is to determine activity. Ideally this should depend only on examining previous chars, already digested by the parser. But since Mathematica allows implicit multiplication, a postscan might be required. Eg sin (a+b) means sin*(a+b) (active blank) but Sin [a+b] means Sin[a+b] (inactive blank)