[Maxima-bugs] [ maxima-Bugs-635627 ] subst([...] is order-dependent

[SourceForge.net <no...@so...>]

Bugs item #635627, was opened at 2002-11-08 11:35
Message generated for change (Settings changed) made by robert_dodier
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=635627&group_id=4933

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
>Category: Lisp Core
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Stavros Macrakis (macrakis)
Assigned to: Nobody/Anonymous (nobody)
Summary: subst([...] is order-dependent

Initial Comment:
I would have expected subst([a=...,b=...],...) to 
substitute a and b simultaneously (like Lisp Sublis), but 
it does not, and it simplifies along the way.  Here are 
some cases where it matters.

The most obvious case is 

   subst([b=c,a=b],b) => c

This means that subst cannot be used to permute 
variables, e.g.

    subst([x=y,y=x],...)

That is not good.

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

But there are other cases:

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

   subst([a=0,b=0],atan2(a,b))

Depending on the answers to a>0 etc., this can return 0 
or %pi, whereas subst([b=0,a=0],...) can return pi/2 or -
pi/2.  I believe that it should give the error: atan2(0.0) 
has been generated.

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

   subst(["="="+","["="*"],[x=1,x=2]);

gives (x+1) * (x+2) as expected, but

   subst(["["="*","="="+"],[x=1,x=2]);

gives x^2+2

((It would have been nice if minus were nary, so that I 
could use "="="-"...))

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

These two cases can be worked around by turning simp 
off temporarily, e.g. 

   subst(["["="*","="="+"],[x=1,x=2]), simp:false;

but the workaround for the first case is much clumsier:

    subst([x=x0,y=x,x0=y],...)


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

Comment By: Stavros Macrakis (macrakis)
Date: 2003-07-12 14:23

Message:
Logged In: YES 
user_id=588346

Another case where simultaneous substitution is necessary...

   subst([a=[1,2],b=[3,4]],a+b) => [[4,5],[5,6]]

instead of the expected [4,6] because [1,2]+b => [1+b,2+b], 
then 1+[3,4] => = [4,5] etc.

I am arguing that simultaneous substitution is the only 
reasonable default behavior.


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

Comment By: Stavros Macrakis (macrakis)
Date: 2002-11-19 12:48

Message:
Logged In: YES 
user_id=588346

Sublis is broken for operators, e.g.

     sublis(["+"="*"],x+y)

And the online documentation (describe) of subst does not 
mention the parallel substitution issue.  I do not think we 
need both subst([...]) and sublis([...]...).  My guess is that 
sublis was defined before subst was extended to cover the 
multiple substitution case.

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

Comment By: Nobody/Anonymous (nobody)
Date: 2002-11-16 09:43

Message:
Logged In: NO 

The info tells you to use SUBLIS if you want to do
substitution in parallel.

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

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=635627&group_id=4933