Run a program

Run a program

[Roland Thiers]

Hi all,
I am a novice programming in Emacs Lisp.
My question :
I dit this program :
(defun coins (n)
(let ((i n) (repartition [0 0 0]) (c1 0) (c2 0) (r 0))
(while (> i 0)
(setq c1 (random 2))
(setq c2 (random 2))
(setq r (+ c1 c2))
(aset repartition r (1+ (aref repartition r)))
(setq i (- i 1)))
(message " repartition %s " repartition)))

aim : you toss two coins, count tails : 0 or 1 or 2. Repeat 10 times  
or more and see what happens.

If I evaluate this program with C-x C-e and do (coins 10) C-x C-e it  
works fine, I get some thing like
[3 5 2].
But if I repeat C-x C-e after (coins 10) I get some thing like [5 11  
4] then [9 15 6] ...
I can't understand that !
Could some one help me ?
Roland Thiers
Noumea New Caledonia
roland dot thiers at canl dot nc

Re: Run a program

[Deniz Dogan]

On 2011-09-27 07:28, Roland Thiers wrote:
> Hi all,
> I am a novice programming in Emacs Lisp.
> My question :
> I dit this program :
> (defun coins (n)
> (let ((i n) (repartition [0 0 0]) (c1 0) (c2 0) (r 0))
> (while (> i 0)
> (setq c1 (random 2))
> (setq c2 (random 2))
> (setq r (+ c1 c2))
> (aset repartition r (1+ (aref repartition r)))
> (setq i (- i 1)))
> (message " repartition %s " repartition)))
>
> aim : you toss two coins, count tails : 0 or 1 or 2. Repeat 10 times or
> more and see what happens.
>
> If I evaluate this program with C-x C-e and do (coins 10) C-x C-e it
> works fine, I get some thing like
> [3 5 2].
> But if I repeat C-x C-e after (coins 10) I get some thing like [5 11 4]
> then [9 15 6] ...
> I can't understand that !
> Could some one help me ?

I am not exactly sure why that happens, but it seems like you're 
referencing the same vector after each call.  To make sure the vector is 
fresh before each run, replace:

(repartition [0 0 0])

with:

(repartition (make-vector 3 0))

Re: Run a program

[Buchs, Kevin]

Roland, 

Please consider using some indentation in your code postings, as we are left swimming in the parens if you don't. 

It is also inexplicable by me that repartition is maintaining a value, despite the let form. It is not recognized globally. You can see that with the additional message form I inserted. Then in addition, I did explicit initialization of each element.

(defun coins (n)
   (let 
      (  (i n) 
         (repartition [0 0 0])
         (c1 0)
         (c2 0)
         (r 0)) 
  
      (message "repartition starts at: %s " repartition)
      (aset repartition 0 0)
      (aset repartition 1 0)
      (aset repartition 2 0)
      (while (> i 0) 
          (setq c1 (random 2)) 
          (setq c2 (random 2))
          (setq r (+ c1 c2))
          (aset repartition r (1+ (aref repartition r))) 
          (setq i (- i 1))) 
      (message " repartition %s " repartition)))

Kevin Buchs   |  Senior Engineer  |  Department of Physiology and Biomedical Engineering - SPPDG
507-538-5459  |   buchs.kevin@mayo.edu  |  http://www.mayo.edu/sppdg
Mayo Clinic  |  200 1st St. SW  |  Rochester, MN 55905  


> (defun coins (n)
> (let ((i n) (repartition [0 0 0]) (c1 0) (c2 0) (r 0)) (while (> i 0) 
> (setq c1 (random 2)) (setq c2 (random 2)) (setq r (+ c1 c2)) (aset 
> repartition r (1+ (aref repartition r))) (setq i (- i 1))) (message " 
> repartition %s " repartition)))


I am not exactly sure why that happens, but it seems like you're referencing the same vector after each call.  To make sure the vector is fresh before each run, replace:
(repartition [0 0 0])
with:
(repartition (make-vector 3 0))

Re: Run a program

[Le Wang]

On Tue, Sep 27, 2011 at 1:28 PM, Roland Thiers <roland.thiers@canl.nc> wrote:
> Hi all,
> I am a novice programming in Emacs Lisp.
> My question :

This is a quoting issue.  It got me a short while ago.  Read the whole
thread, especially when I reach enlightenment later on.
http://lists.gnu.org/archive/html/bug-gnu-emacs/2011-09/msg00457.html

TLDR; instead of using [x y z] use (vector x y z).  Never modify
things you quote or you will be confused.  Only quote constants.

-- 
Le

Re: Run a program

[Roland Thiers]

Thank very much Deniz, Kevin and Le for your replies and advices.
(vector 0 0 0) and (make-vector 3 0) work well.
I understood that my problem was connected with quoting, now I have to  
read  carefully
the thread : buffer-local variables seem to remember values  
(explanations of Drew Adams and others). Not obvious, and my poor  
english
does not help me !
Roland
Le 28 sept. 11 à 12:43, Le Wang a écrit :

> On Tue, Sep 27, 2011 at 1:28 PM, Roland Thiers  
> <roland.thiers@canl.nc> wrote:
>> Hi all,
>> I am a novice programming in Emacs Lisp.
>> My question :
>
> This is a quoting issue.  It got me a short while ago.  Read the whole
> thread, especially when I reach enlightenment later on.
> http://lists.gnu.org/archive/html/bug-gnu-emacs/2011-09/msg00457.html
>
> TLDR; instead of using [x y z] use (vector x y z).  Never modify
> things you quote or you will be confused.  Only quote constants.
>
> -- 
> Le

RE: Run a program

[Drew Adams]

> Thank very much Deniz, Kevin and Le for your replies and advices.
> (vector 0 0 0) and (make-vector 3 0) work well.
> I understood that my problem was connected with quoting, now 
> I have to read  carefully the thread : buffer-local variables
> seem to remember values (explanations of Drew Adams and others).
> Not obvious,

No, it's not obvious to someone new to Lisp.

To be clear, Le took a shortcut in saying that this has to do with quoting.  In
fact, it does not.

What it comes down to is that the Lisp _reader_ constructs Lisp objects as it
reads sexps (code).  That includes lists, vectors, and symbols.  Nearly every
Lisp sexp you see in a program leads to the Lisp reader constructing a list of
lists, symbols, vectors, strings, etc.  (defun foo ...) becomes a list when it
is read, and so on.

Reading is a separate step from evaluating.  Lisp interpretation is a loop with
these steps: Read (a sexp), Evaluate it, Print (the result) - the so-called REP
Loop, or REPL.  People sometimes gloss over the Read step when thinking about
code, but that can be a mistake, as the gotcha in question shows.
http://en.wikipedia.org/wiki/Read-eval-print_loop

A special form such as `quote' does not receive program text as its argument; it
receives an internal Lisp object already created by the Lisp reader.

Depending on the particular Lisp and its implementation, the reader might or
might not create a new, different list or vector each time it encounters
equivalent text (sexp).  In Emacs Lisp, it typically (always?) does not create a
new one - it reuses the same list or vector.

And that is why you should not modify such a "constant", at least not expecting
a new, distinct constant to have been created at each textual occurrence of the
same sexp.  Those textual occurrences will typically all stand for the same Lisp
object.

This is a gotcha associated with learning Lisp, but once it is learned you have
a better understanding of the kind of language it is.

HTH.