On Thu, Feb 19, 2015 at 5:35 AM, Petr Baudis <[email protected]> wrote:
> On Wed, Feb 18, 2015 at 11:01:54AM -0600, Aaron Meurer wrote:
>> On Tue, Feb 17, 2015 at 10:41 PM, Petr Baudis <[email protected]> wrote:
>> > So one curious observation I have is that even if sym.solve() is given
>> > multiple equations and only a single variable, it does yield a solution
>> > in this case.  I'm still confused about what the exact API contract of
>> > sym.solve() is.  Working "sometimes" leaves me unsure about what I can
>> > or cannot rely on...
>>
>> Sadly, there is none. Part of this is due to algorithmic limitations
>> (solve() is basically a bunch of heuristics, there are few guarantees
>> that it will find a solution if it exists). But a lot of it is just
>> poor design. We are trying to make the design better with the new
>> solveset module.
>
>   Thanks!  Knowing this is quite valuable for me.
>
>> >> Right now I'm inclined to conclude that what I want to do *is* rather
>> >> hard and I'd have to come up with some new algorithms to deal with this.
>> >> (Or likely take a look at some other CASes first if any handle this case
>> >> for me already.)
>>
>> If you find some other CAS that can do this better, please let us know
>> and post here your solution. As Aaron said, we would like to improve
>> our solver module.
>
>   I will for sure - my plan is to take a hard look at Maple and a brief
> look at Mathematica.  If I won't be successful, I might come back to
> SymPy and try to help improve the solveset module (I can't promise
> I will have the time though, this was supposed to be just a small piece
> of puzzle in a larger project - http://21robot.org/).

Are you trying to do all of this:

http://21robot.org/research_activities/math/

? If so, that's a *major* project. There is Wolfram Alpha
(http://www.wolframalpha.com/), that has some understanding
of a human language, e.g.:

http://www.wolframalpha.com/input/?i=integrate+cos^2x+from+0+to+2pi&lk=4&num=1

But in my experience once things get a little more complicated, it
can't understand it either, e.g.:

http://www.wolframalpha.com/input/?i=Solve%5B{v%5Bt%5D+%3D%3D+x0%2Ft%2C+a+%3D%3D+v%5Bt%5D%2F%282*t%29}%2C+{v%5Bt%5D%2C+t}%5D

The same expression in Mathematica returns the correct answer (as in SymPy).

Ondrej

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sympy.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sympy/CADDwiVBCSrD4YSEivt7JTQpwLzpB%3DJpO1Crpykfx_yytH05jcg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to