side-effect when evaluating a function

[pca]

Dear all,

I have been puzzled, and annoyed, by an unexpected side-effect. It
can be demonstrated in the notebook by the following example. It
models the speed as distance / time, then solve for the distance, then
evaluate the distance with different arguments:

var('v d t')
equation = v == d / t
solution = solve(equation, d, solution_dict = True)
d_ = solution[0][d]; print d_ #-> t*v
constants= {v: 2}
print d_(constants, t=3) #--> 6
# BEWARE: the previous call has the side effect that t=3 in later
evaluation !
print d_(constants) #--> 6 : the side-effect is apparent here !
print d_ #--> t*v

I would expect d_(constants) to result in t*2, but instead its using a
previous assignment of t to yield 6 !

I'm using Sage for an engineering pblm, and I find it useful to
separate constants from arguments. The side effect is thus annoying
to me: is there any other way to evaluate a function with constants
and arguments, without any side-effect with the arguments ? I'm new
to this forum, so not sure if it has been raised before. Sorry if
there is an easy answer.

Thanks in advance.
PC

[pca]

Thank you for your quick answer, Minh.

Although it would give the correct answer, it does not satisfy my wish
to distinguish constants from parameters. In my engineering pblm, I
have a set of several constants that I use in many statements, and I
wish to keep them together nicely, and use them in only one keyword.
In your solution, I would have to repeat them every time, and this is
not convenient.

Any idea why Sage behaves as it does, and how to circument this
strange behavior while keeping the "constant" idea ?

Thanks
PC

On 25 juil, 21:43, Minh Nguyen <nguyenmi...@gmail.com> wrote:
> Hi Pierre,

> Is the following what you're looking for?
>
> ----------------------------------------------------------------------
> | Sage Version 4.1, Release Date: 2009-07-09                         |
> | Type notebook() for the GUI, and license() for information.        |
> ----------------------------------------------------------------------
> sage: var("v, d, t");
> sage: equation = v == d/t
> sage: solution = solve(equation, d, solution_dict=True)
> sage: solution
> [{d: t*v}]
> sage: sol_d = solution[0][d]; sol_d
> t*v
> sage: sol_d.substitute(v=2)
> 2*t
> sage: sol_d.substitute(t=3)
> 3*v
> sage: sol_d.substitute(t=3, v=2)
> 6
> sage: sol_d
> t*v
>
> --
> Regards
> Minh Van Nguyen

[William Stein]

On Sat, Jul 25, 2009 at 12:31 PM, pca<pierre.ca...@gmail.com> wrote:
>
> Dear all,
>
> I have been puzzled, and annoyed, by an unexpected side-effect.  It
> can be demonstrated in the notebook by the following example.  It
> models the speed as distance / time, then solve for the distance, then
> evaluate the distance with different arguments:
>
> var('v d t')
> equation  = v == d / t
> solution = solve(equation, d, solution_dict = True)
> d_   = solution[0][d]; print d_ #-> t*v
> constants= {v: 2}
> print d_(constants, t=3) #--> 6
> # BEWARE: the previous call has the side effect that t=3 in later
> evaluation !

This is *definitely* a bug in Sage. I've made this trac #6622:

http://trac.sagemath.org/sage_trac/ticket/6622

Here's how to workaround it (namely, put "dict(constants)" instead of
"constants"):

sage: var('v t')
sage: f = v*t
sage: s = {v: 2}
sage: f(dict(s), t=3)
6
sage: s
{v: 2}

William

> print d_(constants) #--> 6 : the side-effect is apparent here !
> print d_ #--> t*v
>
> I would expect d_(constants) to result in t*2, but instead its using a
> previous assignment of t to yield 6 !
>
> I'm using Sage for an engineering pblm, and I find it useful to
> separate constants from arguments.  The side effect is thus annoying
> to me: is there any other way to evaluate a function with constants
> and arguments, without any side-effect with the arguments ?  I'm new
> to this forum, so not sure if it has been raised before. Sorry if
> there is an easy answer.
>
> Thanks in advance.
> PC
>
> >
>

--
William Stein
Associate Professor of Mathematics
University of Washington
http://wstein.org

[pca]

GREAT ! Thanks for the amazing support !

I have just used Sage for one day, and I was disappointed by this side-
effect. But with such an outstanding support, I'm now convinced it's
the tool to use !

PC

On 25 juil, 22:07, William Stein <wst...@gmail.com> wrote: