Bug report ID:635627 - subst([...] is order-dependent

On 4/1/10 11:44 AM, Dieter Kaiser wrote:
> Am Mittwoch, den 31.03.2010, 14:10 -0400 schrieb Joseph Cusumano:
>   
>> Hello:
>>
>> I admit up front to not having the perspective (or skills) of a
>> developer, so maybe I'm missing something fundamental. However, I do
>> not understand why this should be considered a bug. I have a fair
>> amount of code in Maple (which has a similar property to its
>> substitution command) and now Maxima that depends on (or, at the vary
>> least, assumes) this feature property.
>>
>> I would say at the very least this should be controllable via an
>> external variable, e.g. "subst_sequence" equals "serial" or
>> "parallel".
>>     
> Hello Joseph,
>
> thank you for your comment.
>
> Yes, I think it is no problem to get both the parallel substitution and
> the serial substitution we already have.
>
> Because, now the serial substitution is the default and the code of
> to_poly_solver depends on this behavior we might first introduce the
> parallel substitution as an option to the function $substitute.
>   
I didn't follow this dicussion too closely, but why not just leave subst
to do serial substitution, and add, say, psubst to do parallel
substitution? 

I think it's bad to change subst to do parallel because that changes
existing behavior.  There could be lots of code out there that depends
on this and we should gratuitiously break them.  I know this is the
opposite of what Robert said, but I don't see a good reason to break
other people's code in this case.

Ray