% Date: Wed, 22 Jun 1994 15:59:04 +0200
% From: A.J.Tibben <a.j.tibben@math.utwente.nl>

% Thank you for your fast response regarding my problem with factorizing
% with
on complex;
% however another problem has occured. Some simple examples:

factorize((a-1)*(a-b-i));

% doesn't factorize and more seriously

factorize((a-1)*(a-i)*(a-b));

% again leads to a loop.

% I have traced the code a bit and think the problem is in 'gifactor:'.
% The loop there changes 'mvu' such that it doesn't have real roots in
% that parameter, but it checks it for 'mvar u' in sqfrp. I have patched
% it by changing 'mvu:=caar mvu' in 'mvu:=mvar u' and it works up till
% now, but probably the right way to it is by changing the kernel order.

end;

Comment from H. Melenk:

Richard's SUM-bug has a deeper background: k is entered by SUM as formal
argument and SIMPDF cannot see that it is intended as differentiation
order and not as differentiation variable. Because of its syntax
REDUCE is generally unable to operate with differentials of undetermined
order.

REDUCE Development Version, 14-Jun-94 ...

1: df(sin x,x,k);

0
