See http://keiapl.info/rhui/remember.htm at the paragraph that begins: Ken and I had in mind to implement A Dictionary of APL Several attempts were made to implement fork as an operator or operators. As for bidents, see http://www.jsoftware.com/jwiki/Essays/Hook_Conjunction?
----- Original Message ----- From: Raul Miller <[EMAIL PROTECTED]> Date: Monday, July 2, 2007 12:22 Subject: [Jchat] A mad idea To: Chat forum <[email protected]> > What would J be like if hooks and forks were not an artifact of > the parser > but were implemented as explicit commands? > > Fork would be an adverb which takes a verb and produces a conjunction. > The result of that conjunction would be a verb. > > Hook would be a conjunction. > > Trident would disappear from the parsing rules. > > Bident would have semantics like U @: V > > This would make some expressions much easier. Expressions which > chain monads together would become much simpler. However, > expressionswhich chain dyads together would be a bit more > complicated -- counting > from the right you would need to decorate all the even verbs > with the > fork conjunction. > > For example, if fork were ]: the classic J expression for mean would > become: > mean =: +/ %]: # > > The dictionary would need to be changed. Every book or > page ever > written about J would need to be changed. > > Worse, implicit parenthesis involving forks would follow different > rules from implicit parenthesis involving non-forks. To > work around > this, you'd have to rewrite the rules for adjectives and conjunctions. > > The result would not look much like J. For example, the classic > J expression for mean might be more like > mean=: /+ %]: # ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
