4th.CoSy update : "atomic" apply .

Even doing nothing , I haven't had blocks of time to make much
progress on 4th(reva).CoSy since APL2007 .  I just , however ,
uploaded a new copy with a corrected version of my "atomic" apply
"adverb" . Below is the msg I've posted both on the Reva Forth forum
at http://ronware.org/reva/viewtopic.php?pid=5993#p5993
and on the
CoSy forum at http://www.cosy.com/Forum/viewtopic.php?pid=28#p28
where
it has the advantage of better formatting . Anybody is welcome to
download and "join" CoSy development at any time , of course , but
it's better to think of it as extensions to Forth than a full APL at
this time .

4th.CoSy at this level executes directly in Forth which means each
word ( chars delimited by whitespace ) is interpreted as
encountered .

| ================================================== |

I've uploaded updated files to http://cosy.com/CoSy/CoSy/CoSy.zip
.
Not many changes . I'll wait for float x^y etc for the integrated Reva
release .

Actually , I uploaded a new copy mainly to be sure to support a few
lines ( below ) I composed without crashing once . ( Did a ctrl_s
"save" virtually every step tho . )

I decided I wanted to get a full list of all the days of the year .  I
ended up thinking it's not very useful . I really would like to pull
up the brilliant calendar formatting APL from Gene McDonnell I used in
2nd.CoSy .

You will see I added a new word ^+ which will be the "CoSy" level
definition of + . Right now it's just integer add . This would be
called an atomic function in APL . The operator or adverb aaply
applies the ticked function over the corresponding leaves of LA and
RA . Remember indexing in CoSy is modulo the length of the longer list
argument .

Because this is such a quintessential "APL"ish function , and shows
what a major generalization and simplification modulo indexing is ,
here are a number of examples with results of aaply on simple arrays .

Code:

: ^+ ( LA RA -- itemwise_LA+RA )  ['] + aaply ;    | "atomic" apply
integer +  | applies fn between leaves of LA and RA

 i( 1 2 3 )i dup ^+ lst   | sum of vector with itself
2 4 6

  12 _iota i( 3 4 )i reshape >t0> lst   | A 3 item list of 4 atom
lists
(
 0 1 2 3
 4 5 6 7
 8 9 10 11
 )

  i( 10 )i t0  ^+ lst   | APL's "scalar extension"
(
 10 11 12 13
 14 15 16 17
 18 19 20 21
 )

  t0 dup ^+ lst   | "matrix" addition
(
 0 2 4 6
 8 10 12 14
 16 18 20 22
 )

  i( 10 20 30 )i t0  ^+ lst   | Vector + matrix
(
 10 11 12 13
 24 25 26 27
 38 39 40 41
 )

  i( 10 20 30 40 )i t0 ' ^+ eachright lst   |
(
 10 21 32 43
 14 25 36 47
 18 29 40 51
 )

  12 _iota i( 3 2 2 )i reshape >t0> lst        | A 3 item list of 4
atom lists
(
 (
  0 1
  2 3
 ) (
  4 5
  6 7
 ) (
  8 9
  10 11
 ) )

  i( 10 20 )i t0 ^+ lst
(
 (
  10 11
  12 13
 ) (
  24 25
  26 27
 ) (
  18 19
  20 21
 ) )

| Some other examples to try :
  i( 10 20 30 )i enc enc t0 ^+  lst
  i( 10 20 30 )i ' _iv eachM> lst
  i( 10 20 )i i( 20 )i ,l i( 30 )i cL t0 ^+ lst

Now to get to the original ( pointless ) calendar example which
motivated all this .  Download CoSy if you want to dissect each line .
Code:


  1 1 2008 date>fixed _iv  365 _iota ^+   | Convert New Years to
"julian" date , make into CoSy vec , add 365 days
  { fixed>date dtpk } eachMcr fmtI   | convert each back to
"gregorian" dates and pack as yyyymmdd integers , then format as
strings .
  4 _iv ' cut eachleft  ' cL across lst   | Lop off "yyyy" from each ,
list
(
 s" 0101"
 s" 0102"
 s" 0103"
 s" 0104"
 s" 0105"
 s" 0106"
 ...  )