You're perfectly right, Tom. J documentation does have its blind-spots... With my mathematician brain plugged-in, I'm contemptuous of mortals who can't grasp the blinding obviousness of J in its ultra-consistent totality, and its instantiation to any conceivable goal. ;-)
With my applied-psychology brain plugged-in, it's clear to me there's a dearth of task-oriented documentation. It's nearly all function-oriented. In other words, you have to know the function you need in order to look it up. That being so, people who've learned J successfully have done it by reading the whole of the J Dictionary, on 2 tablets of stone (Dic/Voc), making notes on the way. In other words, writing their own task-oriented documentation. I still do that, in the few spare moments I snatch from tussling with J to make it do what I want. If you come from an APL background (APL and cousins, I mean: you mentioned K and A+) then take a look at: http://www.jsoftware.com/jwiki/APL2JPhraseBook It is precisely oriented to one particular task: find the J construct to look up in order to perform the task that APL offers (say) {branch} (→) to achieve. (As it happens, you'll find "return." there.) If you come from a C background, then Henry Rich's "J for C programmers" plugs a similar gap. Another thing about the body of J literature is that it's mainly in Tacit Code. So... no control structures. For good samples of Explicit Code, you can't do better than explore the contents of the system locales: _z_, _j_, _jijs_ . The J wiki describes two code browsers to help you do exactly that. Plus the demos and labs, which will answer nearly all your task-oriented questions without tears. J602 is better at delivering the Labs than j701. The other way is to carry on doing what you did: ask in the programming forum (better than this one: the "general forum", which is supposed to be for J installation questions.) We're very supportive, because we all know it's the only real way to find out anything. On Wed, Oct 17, 2012 at 4:06 PM, Tom Szczesny <[email protected]> wrote: > Also (interestingly): > 1) There is no entry for "return." in the index. > 2) Until I got the your responses to my email, I could not find a > relevant entry in > forum search for "return" > > It might be just a further manifestation of my dyslexia, but I find it > somewhat > difficult to find answers to specific questions in the J help documents. > > Anyway, thanks again for your quick response. > > On Wed, Oct 17, 2012 at 10:54 AM, Tom Szczesny <[email protected]> wrote: >> Thanks. >> I was looking for a monadic primitive verb, >> but of course, in J, it is a control structure. >> (It is not listed as one of the control structures in the Primer). >> >> On Wed, Oct 17, 2012 at 10:06 AM, Ian Clark <[email protected]> wrote: >>> How about: >>> 99 return. >>> >>> corresponds (roughly) to (APL): >>> >>> ∇z←a foo b >>> ... >>> →0,z←99 >>> ... >>> >>> >>> On Wed, Oct 17, 2012 at 2:11 PM, <[email protected]> wrote: >>>> In some APL-related languages there exists a nomadic verb for "early >>>> return" from a defined function. In A+ this is "result" (nomadic left >>>> arrow). In k this is nomadic colon. Does a similar capability exist in J? >>>> >>>> Sent from my iPhone >>>> ---------------------------------------------------------------------- >>>> For information about J forums see http://www.jsoftware.com/forums.htm >>> ---------------------------------------------------------------------- >>> For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
