Oh well, I should have remembered that. I even replied to it. :-)
On 8 February 2016 at 00:17, Juergen Sauermann < juergen.sauerm...@t-online.de> wrote: > Hi Elias, > > great idea, except that it conflicts with the NARS APL interpreter which > had defined > Rank with axis already (and, of course, differently). > > See http://lists.gnu.org/archive/html/bug-apl/2014-07/msg00042.html > > /// Jürgen > > > On 02/07/2016 01:31 PM, Elias Mårtenson wrote: > > How about implementing a non-conflicting extension? I'm suggesting the > following form as an alternative: > > Z ← f ⍤[j] B > Z ← A f ⍤[j] B > > Regards, > Elias > > > On 7 February 2016 at 20:27, Juergen Sauermann < > juergen.sauerm...@t-online.de> wrote: > >> Hi Alex, >> >> not sure what *a ¯3⍤< m* is supposed to mean. >> >> According to the ISO standard the syntax for *⍤* is: >> >> *Z ← f ⍤ j B *(monadic, page 124) or >> *Z ← A f ⍤ j B *(dyadic, page 125) >> >> If you compare that with your example: >> >> a ¯3 ⍤ < m >> >> then the (expected value) *j* is the primitive function *<*, which >> triggers the *VALUE ERROR*. >> The fact that the caret points to *a* is not because *a* is the culprit, >> but because *a* is the left >> end of the phrase being reduced. >> >> Unfortunately the syntax in the ISO standard is somewhat ambiguous:* j *is >> a one, two, or three >> element vector, and B is the rest. Therefore it is sometimes impossible >> to decide where *j* ends >> and where *B* begins, and the examples for *⍤* in the ISO standard are >> in conflict with the IBM APL2 >> binding rules. This conflict occurs only with *⍤* which - wise decision >> - is not implemented at all >> in IBM APL2. >> >> The conflict can be avoided by always putting *j* and *B *into separate >> variables. If you use literals for *j* >> or *B*, heaven forbid, then be prepared for fairly nasty error messages >> at times. >> >> /// Jürgen >> >> >> >> On 02/06/2016 05:17 PM, alexwei...@alexweiner.com wrote: >> >> >> Hi bug-apl, >> >> Why am I getting a value error here? It seems that the variable 'a' >> definitely exists: >> >> a←2 2 ⍴2 >> m←4 4 ⍴⍳16 >> a >> 2 2 >> 2 2 >> m >> 1 2 3 4 >> 5 6 7 8 >> 9 10 11 12 >> 13 14 15 16 >> >> >> >> VALUE ERROR >> a ¯3⍤<m >> ^ >> a >> 2 2 >> 2 2 >> >> >> >> >> SVN 693 >> -Alex >> >> >> > >
