Hi,
>> Somebody should check (re-check ?) branch cuts and singularities with this
>> patch.
>>     
>
> (for convenience, I include the patch below)
>
> I guess the right place to check branch cuts and singularities is in
> src/input/elemnum.input.  I looked at the common lisp hyperspec, the 
> definition
> of branchcuts is quite clear there, so I suggest that we take it as 
> definition,
> unless you have a better reference.
>   

I don't understand.

Do you change asin z == atan (z/(1-z^2)^(1/2))
to                       asin z == -i log(iz+(1-z^2)^(1/2)) ?

I prefer the atan definition because real number remains real number.

> so we need to test:
>
> real     < -1     error
>          > +1     error
>          = -1     -%pi/2
>          = +1     -%pi/2
>   
Yes this test is necessary for the atan definition, with the denominator
(1-z^2)^(1/2).
No this test isn't useful for the log definition : log (e*%i) == e*%pi / 
2 almost everywhere for e=1 and e=-1.

> complex  x + 0*%i, x < -1 close to x + 0.1*%i
> complex  x + 0*%i, x >  1 close to x - 0.1*%i
>
> Is this correct?
What definition do you call ?
with the log definition I see no sigularity arroud x+/-0%i for x >=1.

x^2-1 is (c+) +/- 0%i : a positive constant (c+), and imaginary part is 
arround 0.
sqrt(1-x^2) is arround (c+)*%i
then (%i*z+sqrt(1-z^2)) = (c+)*%i
and -i log (iz+sqrt(1-z^2)) = pi/2 +/- %i*0...

If asin is defined from the complex atan, what definition do you use for 
atan ? I suppose you start from :
http://www.lispworks.com/documentation/HyperSpec/Body/f_asin_.htm
 
i (log (1-iz)-log(1+iz)) / 2 = -i log ((1+iz)/(1+z^2)^(1/2))

François

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"FriCAS - computer algebra system" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/fricas-devel?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to