There is another argument mentioned in the essay http://www.jsoftware.com/jwiki/Essays/Hook_Conjunction%3F namely that the hook (along with other bidents) required a special parsing rule. A hook conjunction would not. Since there are only 9 parsing rules, a reduction by 1 is significant.
All of this "regrets" discussion is academic. None of it is going to be changed. ----- Original Message ----- From: Raul Miller <[EMAIL PROTECTED]> Date: Friday, December 21, 2007 11:27 Subject: Re: Re[Jgeneral] grets To: General forum <[email protected]> > On Dec 21, 2007 2:03 PM, Dan Bron <[EMAIL PROTECTED]> wrote: > > My objection to the hook conjunction is that it's often uglier > than the > > current bident. Further, I do not see the value of > providing that > > equivalence. Do you have a compelling example? > > I am not concerned with "ugly". I am not even sure what > "ugly" means. > > I am concerned with legibility. > > Currently, we often have to represent hooks as (f (g)) because g is > a train and counting long trains to determine syntax is not nice. > > That said, the backwards compatability argument (lots of existing > J code uses bident hooks) is a compelling argument against any > change in bident. > > I was just wondering if there were other compelling arguments > that I could understand. > > -- > Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
