Porting Postscript code to TRANSCRIPT

[Mark Wieder]

Alejandro-

Saturday, January 17, 2004, 1:55:40 PM, you wrote:

AT> But I do not understand how these commands or
AT> functions works on the data: dup store exch mat sub.

Well, this is based on forth, not postscript, but:

dup: takes the top item on the stack, makes a copy and leaves both
copies on the stack. If you say
1 2 3 dup
you will leave this on the stack:
1 2 3 3

store: do you mean grestore? I don't know what store would be.

exch: in forth this is call swap. It takes the top two items on the
stack and reverses their position. If you say
1 2 3 exch
you will be left with
1 3 2

mat: this looks like a command defined in the code. See the line
/mat {rawdata} def
which looks to me as if it takes the matrix of raw data and stuffs it
onto the stack.

sub: subtract.
length 1 sub will leave length-1 on the top of the stack.

--------------

So the line
/finderrdist {ytt sub exch xtt sub dup mul exch dup mul add sqrt} def

creates a handler called "finderrdist" that returns the square root of
(x squared + y squared), where x squared is the square of the sum of
the difference between the xvalue passed to finderrdist and the value
xtt, and y squared is the difference between the yvalue passed to
finderrdist and the value ytt.

Think of it unrolled as:
ytt sub     -- y - ytt
exch        -- get x out from under the result
xtt sub     -- x - ytt
dup mul     -- x^2
exch        -- get y out from under x squared
dup mul     -- y^2
add         -- x^2 + y^2
sqrt        -- square root of the result

This comes out in Transcript as:

local xtt
local ytt

-- Return the error distribution of a point x, y
-- from the point xtt, ytt
function finderrdist x, y
  local yErr
  local xErr
   
  put x - xtt into xErr
  put y - ytt into yErr
  return sqrt(xErr^2 + yErr^2)
end finderrdist

-- 
-Mark Wieder
 mwieder at ahsoftware.net