I have added code to handle real-valued functions and functions declared to be
real-valued for the routine rectform.
signum simplifies as a real function:
(%i10) rectform(signum(x));
(%o10) signum(x)
A unknown function:
(%i12) rectform(f(x));
(%o12) 'realpart(f(x))+%i*'imagpart(f(x))
The function can declared to be real-valued:
(%i13) declare(f,real);
(%o13) done
(%i14) rectform(f(x));
(%o14) f(x)
A function which is declared to be integer-valued will simplify too:
(%i15) declare(n,integer);
(%o15) done
(%i16) rectform(n(x));
(%o16) n(x)
All this complex-, real-, integer-valued stuff and properties can be made much
more consistent and complete. Further work is needed.
Dieter Kaiser
-----Urspr?ngliche Nachricht-----
Von: Barton Willis [mailto:willisb at unk.edu]
Gesendet: Samstag, 9. Mai 2009 02:10
An: Dieter Kaiser
Betreff: Re: [Maxima] Declare signum(x) to be real-valued
Thanks for checking on this.
Barton
-----maxima-bounces at math.utexas.edu wrote: -----
>Because?of?the?last?commit?to?rpart.lisp?which?changes?the?standard
>behavior?of?rectform,?the?signum?function?no?longer?simplifies?as
>expected.?Therefore?we?get?failures?for?the?tests?in
>rtest_abs_integrate.mac.?I?have?overlooked?this?fact.
>
>Tomorrow,?I?will?commit?a?correction?to?declare?signum(x)?a?real-valued
>function,?which?will?simplify?as?expected?when?rectform?is?applied.
>
>Dieter?Kaiser
>
>_______________________________________________
>Maxima?mailing?list
>Maxima at math.utexas.edu
>http://www.math.utexas.edu/mailman/listinfo/maxima