Thanks for your comments Don. Whilst I agree with your desire to improve the notational side of J I think there is an issue with ASCII to do so...
If you wrote a table of all graphic symbols required for each token (word), you would find that the list of ASCII symbols is insufficient to accommodate this. I believe that a key principle of Ken, Eugene and Roger was to stay with ASCII (for accessibility and universality at the time), so there was no choice other than to expand the dictionary tokens to 2 character glyphs. If you can see a simple transliteration table that could be used, then you could model this in J (ie write the "A" option as a J preprocessor yourself). You may achieve some success, but I suspect the outcome will be to realise that it can't be fully achieved without extending the ASCII set. You may end up heading towards a UTF-8 implementation (or other expanded character set) as Don suggested. Rob Hodgkinson On 23/02/2009, at 5:22 AM, Don Watson wrote: > Our university acquired its first APL system in 1968 – after I heard > Ken Iverson give a seminar on “The Role of Computers in Teaching”. > In those days software was free from computer manufacturers. Iverson > promoted APL vigorously as an aid in the teaching of Mathematics and > its applications - and it was so easy to teach - students from many > disciples used it for their classes > > > > I have been retired for 13 years, and haven’t used APL or anything > like it for 25 years. I recently started providing enrichment to a > gifted 10 year old and sought an APL system. I ended up with a J > system instead because APL systems appeared to be: > > > > · Not available under Windows, or > > · Not well supported, or > > · Expensive, and > > a.. Divergent from each other, > b.. With keyboards that cannot serve regular ASCII purposes. > > Although APL systems today are used commercially for manipulating > data, it is my impression that fewer students use APL (or J) in > their courses than 40 years ago. There certainly has not been the > expansion in teaching use that there should have been. This is a > shame. Today’s smaller pool of student learners surely also reduces > the opportunities for commercial sales of APL and J systems. > > > > I am sure that experienced J programmers are used to the language > and have no problems with it; however, J is not easy for (say) a 10 > year old public school student, a 15 year old high school student or > a geography major. J may be as elegant in its structure as the > original APL, but APL had those symbols that were one character > instead of two and expressed meaning through their design. > > > > Can J be turned into be a revised system that better meets the needs > of the teaching community as well as the needs of the commercial > community? Such a J system would need to fit all of the following > criteria: > > > > 1. Acceptable as a standard. > > 2. Free for teaching purposes. > > 3. Powerful. > > 4. Well supported. > > 5. Able to use the same keyboard as regular ASCII applications. > > 6. Available on PCs and Macs. > > 7. With teaching additions: > > a. As easy to read and learn as possible. > > b. Transparent to current J users. > > c. Retaining the ability to use the same keyboard as regular > ASCII applications. > > d. Easy to add to the J system. > > > > These criteria can be met very easily: > > > > Criterion 1 J is a good standard on which to build a new teaching > system, because it has the following advantages: > > > > · J fits Criteria 2 through 6. > > · Of the successors to the original APL, only J had Iverson’s > approval. The alternative, designing changes by committee, according > to commercial interests or by adversarial competition is unlikely to > be in the best interest of the teaching community. > > > > Criterion 7a An addition can be made to the J system to make it more > suitable for teaching, by the replacement of two character symbols > on the screen with single symbols designed to have meaning related > to their function. > > > > Criterion 7b The revised J system could have a “J” mode identical to > the present system and (say) an “A” mode for teaching. The “A” mode > might also be preferred by converts from current APL systems. > > > > Criterion 7c The same keyboard could be used for ASCII applications > if the revised system: > > > > · Required exactly the same keystrokes in the “A” mode as in > the “J” mode. > > · Placed all three related symbols on the same key (e.g., = > with new symbols replacing =. and =:). > > · Had a convention in their positioning as to which one came > from one keystroke, which from adding “.” And which from adding “:”. > > > > Criterion 7d As a result, the system changes in J would be very easy > to make - the only change in the “A” mode would need to be in > looking in a two character table to find the single symbol > replacement and presenting it on the screen (and in printing). > > > > Since such a change would be so easy to implement, someone must have > already produced such a system. Could you please own up? I’d like a > copy. > > > > Don Watson > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
