On 07/23 10:39, Raul Miller wrote: > On Sat, Jul 23, 2022 at 9:11 AM <tu...@posteo.de> wrote: > > My preffered way of learning J to the point, where I will be able to do my > > own > > experiments with it, "predict", what I would expect as a result and finally > > find the bugs I did myself by reading online resources (again the dense ones > > and not the "hello world" ones), would be to read sources like the NuVoc and > > https://sergeyqz.github.io/jcheatsheet/ > > This matches my experience. > > And the J labs (jqt menu: Help > Studio > Labs...) are designed to > somewhat cater to this approach, but other resources exist. > > > And I think, a "problem" (no native speaker, sorry) of these sources is, > > that > > they are explaining J in and with J terms. > > I remember also being frustrated with this issue. That said, what's > the alternative? > > Conceptually, we often get a better match with user interests when the > J content is explained within an application domain that the learner > is interested in. But that requires someone to have made the effort to > build out some relevant content in the context of that domain. Some > people have done such things -- for example: fractals visualization > and j by Clifford Reiter -- but there are many areas of potential > interest. > > > No harm meant! ... only my two mico cents... > > You currency may vary... :) > > If you feel motivated to tackle one of these projects, I guess taking > some notes along the way might help later on in refining the effort? > > Thanks, > > -- > Raul > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm
Hi Raul, one addition, which comes to my mind after I sent my post (why is it always "after" :) ) The way, programs (in the sense of "application" not in the sense of "experiments") are written most of the time in other languages (those great bunch of more or less C-like languages) - I think - is: 1) Write class/couple of functions/other greater entities 2) compile (if it is a compiled language) 3) run and check. if BUG!: correct code, GOTO 2) 4) deliver (ok, that's very coarse) With J I got the impression that it works this way: 1) Start the REPL 2) Enter some first steps, try and retry until it seems to do what you want 3) Expand to the next level of <missing word here> (maybe "abstraction" is correct?) 4) GOTO 2) correct some things, change others GOTO 3) 5) Refine between 2) and 4) until your expection and the result of your code match 6) Ready It's more like a "guided experiment" with tends towards the wanted result. Example: The Goldbach conjecture https://yewtu.be/watch?v=Dh4QKfH6fTc ( This is by no means meant as any form of critism towards the author tangentstorm!!! ) Is it possible to write a J application down, start the interpreter and fix a few things? Yes, I would like to try to add content to make learning J easier for others. But this is when the cat bite in its own tail (direct translation from german to english for a "catch-22" situation): I need to learn J up to a level, when I am fluent enough to create such a content. And as with all of us: Time is an additional problem. Let's see. Cheers! Meino ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
