Efficiency is usually not the main concern when generating the identity matrix. You'd run out of space well before the time would matter.
I myself prefer the really short methods, for example, [EMAIL PROTECTED] ----- Original Message ----- From: "R.E. Boss" <[EMAIL PROTECTED]> To: "'Programming forum'" <[email protected]> Sent: Thursday, March 23, 2006 8:23 AM Subject: RE: [Jprogramming] monad e. Your method seems to be equivalent to I3a in http://www.jsoftware.com/jwiki/Essays/Identity_Matrix but nothing is said there about performance. R.E. Boss -----Oorspronkelijk bericht----- Van: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Namens Devon McCormick Verzonden: donderdag 23 maart 2006 17:05 Aan: Programming forum Onderwerp: Re: [Jprogramming] monad e. I use the following to produce an identity matrix because it's faster: idMat=: 2&$ $ ([: >: ]) {. 1"_ which is a tacit version of idMat=: 3 : '(2$y.)$(>:y.){.1' Can't say I've ever used monadic e. but I like Oleg's pictures. On 3/22/06, Henry Rich <[EMAIL PROTECTED]> wrote: > > Except for e. i. n to produce an identity matrix, has anybody > ever used monad e. ? > > Henry Rich ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
