As Eric has pointed out, J has carefully picked related pairs of ASCII characters which graphically show the connection between related functions. APL did this as well. The problem arises when we realize that there are many J primitives in related groups which don't have any APL characters that would fit, and neither are there any sets of unicode glyphs which have the appropriate graphical characteristics that would suggest that relatedness.
It is clear. To do the J-to-single-glyph conversion right we would need some new glyphs. That would likely require an expert graphical designer who was also either a mathematician or a programmer, who could express the functionality AND the relatedness of related primitives in a single glyph. Good luck with that. It could be done by someone with the right skill set, but who (or what group of people) would that be? That brings up an interesting question... How DID the APL character set get designed? Some IBM graphic designers? Ken? Who? Skip Skip Cave Cave Consulting LLC On Thu, Feb 27, 2014 at 4:49 PM, PMA <[email protected]> wrote: > I'm most concerned, that the integrity of J's vocabulary's > internal relationships (cited yesterday by I-forget-whom) > not be compromised. > > P > > > Eric Iverson wrote: > >> It is trivial to display/enter fancy single glyphs for J. At least as >> trivial (which I personally don't consider to be trivial) as it is for >> APL. >> >> Those interested (and I am definitely not) should ignore this trivial >> aspect of the problem and focus on the hard part which is the mapping >> of single unicode glyphs to J primitives. >> >> What are the glyphs for =. =: +. *. *: =: ? ?. etc.? >> >> Perhaps if I saw a complete proposal for this glyph to J mapping that >> had some communitiy consensus, I wouldn't be so profoundly negative on >> discussions like this. >> >> Keep in mind that the APL folk paid enormous attention to the >> appearance of the APL glyphs. A hodge podge of unicode glyphs from >> random parts of unicode fonts would not please anyone who loved APL. >> >> We have beaten this dead horse every year for 25 years. When will it >> be out of its misery? >> >> >> Something that people >> >> On Thu, Feb 27, 2014 at 5:10 PM, Joe Bogner<[email protected]> wrote: >> >>> On Thu, Feb 27, 2014 at 4:56 PM, Skip Cave<[email protected]> >>> wrote: >>> >>>> Just my two cents worth... >>>> >>>> As an old APL (occasional) programmer, I always wanted a way to flip a >>>> switch in the J editor and turn J's 2-character primitives into APL >>>> characters (where appropriate), and either leave J's unique verbs alone, >>>> have the community decide on an appropriate single glyph, or let me >>>> pick a >>>> symbol for those myself. Then I could always flip that switch in the >>>> editor >>>> back, and see the actual J code, any time I wanted. >>>> >>> >>> >>> It seems like it would be straightforward to create a JHS editor and >>> viewer that toggles between the character sets. Could it just parse >>> the words and replace with the APL character from a lookup table? The >>> editor would replace back to ASCII before evaluation. >>> ---------------------------------------------------------------------- >>> 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
