Right, and if a goal is to manipulate the representations of functions to generate other functions (that include tacit adverbs in addition to verbs) it might advisable to use the atomic representation directly which was, loosely speaking, Bob Therriault’s advice to Blake McBride. Regarding adverbial programming the first bug’s bite causes some confusion and subsequent bites only minor irritations.
________________________________________ From: programming-boun...@jsoftware.com [programming-boun...@jsoftware.com] On Behalf Of Roger Hui [rogerhui.can...@gmail.com] Sent: Wednesday, January 04, 2012 12:23 PM To: Programming forum Subject: Re: [Jprogramming] Turning verbs to and from strings There is a workaround, namely applying the atomic representation (5!:1) and then applying the string representation (5!:5) to *that.* (The inverse is execute, and then applying 5!:0 to *that*.) The result however is much less readable in the vast majority of cases. On Wed, Jan 4, 2012 at 9:06 AM, Jose Mario Quintana < josemarioquint...@2bestsystems.com> wrote: > There is also a resistant bug in the linear representation that might be > important to (tacit) adverbial programmers; see > http://www.jsoftware.com/pipermail/general/2009-August/033231.html > > ________________________________________ > From: programming-boun...@jsoftware.com [programming-boun...@jsoftware.com] > On Behalf Of Ian Clark [earthspo...@gmail.com] > Sent: Tuesday, January 03, 2012 10:18 PM > To: Programming forum > Subject: Re: [Jprogramming] Turning verbs to and from strings > > Yes, that's my understanding about 5!:5 too. > > > Am I missing something important? > > Yes you are missing something important ...to me. > > I guess you're puzzled by repnoun in: > > repn=: 3 : 0 > NB. Re-inputtable represn of item (y) > if. 'NB.' -: 3{. y do. y return. end. NB. output a comment as-is > select. t=. >type y > case. 'invalid name' do. y repnerr t > case. 'not defined' do. y repnerr t > case. 'noun' do. repnoun y > case. 'adverb' do. 5!:5 <y > case. 'conjunction' do. 5!:5 <y > case. 'verb' do. 5!:5 <y > case. 'unknown' do. y repnerr t > end. > ) > > repnoun is a sort of plugboard to help me quickly slot-in a special > treatment without need for debugging or side-effects. In the untouched > cases (most of them, in fact) it falls back on: cr -which is just a > cover for 5!:5 . > > Why? > > I generate a lot of scripts with Lobrow, intended for hand-editing, > and I now-and-then encounter cases where 5!:5 gives IMO an unhappy > result, which I fix-up on-the-fly. > > For instance: > > zz=: 'alpha',LF,'bravo',LF,'charlie' > > 5!:5 <'zz' > 97 108 112 104 97 10 98 114 97 118 111 10 99 104 97 114 108 105 101{a. > > But... > > repnoun 'zz' > }: 0 : 0 > alpha > bravo > charlie > ) > > > On Mon, Jan 2, 2012 at 4:50 PM, Raul Miller <rauldmil...@gmail.com> wrote: > > I believe that > > > > name,'=: ',5!:5 <name > > > > should give a definition for any valid name which is suitable for use > > in a script file. > > > > The deficiencies I am aware of for 5!:5 have to do with using its > > result as an argument for (".) > > > > When I looked at your code to see what you were not using 5!:5 for, I > > see that you are not using it for nouns. But I believe that all bugs > > involving 5!:5's treatment of nouns have long since been fixed. > > > > Am I missing something important? > > > > Thanks, > > > > -- > > Raul > > > > > > > > On Sun, Jan 1, 2012 at 7:33 AM, Ian Clark <earthspo...@gmail.com> wrote: > >> If you're writing your own IDE for J, then see sample code at: > >> http://www.jsoftware.com/jwiki/IanClark/LoBrow > >> > >> LoBrow makes a legible char array from not just an explicit verb but > >> any J word. To do that it handles each case separately. See the verb: > >> repn (which mainly uses 5!:5), called by: item2panel. > >> > >> But the reverse operation, broadly comparable to []FX, is done using > >> 0!:111 -which works for everything. Provided you handle your > >> "character matrix" as a LF-separated string. > >> > >> Or 0!:110 if you don't want it to display in the session window. > >> > >> On Sat, Dec 31, 2011 at 3:50 PM, Blake McBride <bl...@mcbride.name> > wrote: > >>> Greetings, > >>> > >>> More than 25 years ago I wrote an IDE for APL. At the core what I did > >>> (which I don't remember how now) is that I executed something to turn > >>> an APL function into a character matrix. I would then edit it with me > >>> IDE editing facility. When I saved the edited function it would turn > >>> the edited character matrix back into a function (was that quad FX?). > >>> > >>> I also used this ability (to turn an APL function to and from a > >>> character matrix) to write programs that generated other functions. > >>> This was very nice. > >>> > >>> This leads me to my question - how can I turn a J verb into a > >>> character matrix, and how can I turn a character matrix back into a > >>> verb? > >>> > >>> Thanks! > >>> > >>> Blake McBride > >>> ---------------------------------------------------------------------- > >>> 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 > > ---------------------------------------------------------------------- > 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
