[aplusdev] Neophyte to A+ tries out New Scientist puzzle

[Martin Sitte]

    How would the group feel about the following suggestion?  a+ code
should be entered as itself -- possibly copied from an editor that uses
the kapl font, if the mail client can't.  It can then be read either by
temporarily selecting kapl as the mail reading font in the mail client,
or by copying the message to a viewer or editor that can use kapl.

    Disadvantage: awkward, and tedious if both mail client and editor
have to be used.
    Advantage:  graphics are slow loading, can't be edited in
responding.

    That said, some suggestions (not having Linux, I can only speak from
memory of a+, and can't test any assertion that follows):

    either , at 0 0 0 or , at 0@0 1 should give outer join
    do not use enclosed arrays unless necessary (in all APLs): less
efficient
    in the absence of enclosure !@2 or r[;0;], at 1 r[;1;] will obviate
,reduce
    r[(/@1 r[;0 1]=r[;3 6])/É#r]
    howmany y:{+/(yÉy)=ÉÒy}
    6= howmany at 1 r    ã no enclosing
    (none of these have been checked...)

    Which is all to say, a+ is built around the rank operator, which
should more often than not be tried first.

    P.S.:  Hi, Brian... Hi, Dan...