sign of expression in integration



On Sun, Mar 29, 2009 at 07:13:25AM -0400, R Fateman wrote:
> Semyon Glazyrin wrote:
>> Hello!
>>
>> When I try to integrate one simple expression maxima asks me about
>> sign of one subexpression. But I think in this case integral limits
>> answer the question:
>>
>> (%i2) assume(a>0,b>0);
>> (%o2)                           [a > 0, b > 0]
>> (%i3) integrate(sqrt(a/x-b),x,0,a/b);
>> Is  b x - a  positive, negative, or zero?
>>
>> n;                                     %pi a
>> (%o3)                              ---------
>>                                    2 sqrt(b)
>>
>> Does maxima do such type of checks?
>>   
>
> sometimes.   if you leave out all assumptions it behaves differently, so it 
> uses them for something.
>
> if you say  assume(b*x-a<0)  then it does not ask the question, and just 
> gives the answer.
>
> Which behavior represents a bug is not entirely clear.   Since x is a 
> global variable in the "assume"
> and x is a (presumably different) bound variable in the integration, should 
> the assumption have
> any effect?  On the other hand, shouldn't the integration program figure 
> more of this out itself?

I think it's a problem of integrate function - it should do checks using boundary
conditions.

Semyon.

> RJF
>
>> thanks,
>> Semyon.
>> _______________________________________________
>> Maxima mailing list
>> Maxima at math.utexas.edu
>> http://www.math.utexas.edu/mailman/listinfo/maxima
>>   
>