Signed infinity and signed zero Re: [Maxima] proposal to change acos

>>>>> "Richard" == Richard Fateman  writes:

    Richard> There are lots of rules, most of which are consistent,
    Richard> and some of which help with branch cuts.

    Richard> For signed stuff, we can use
    Richard> 0/0 = und rational
    Richard> 1/0 = pos inf
    Richard> -1/0 = neg inf
    Richard> 0/-1 = neg zero
    Richard> 0/1 = 0 = pos or unsigned zero

    Richard> at least that was my proposal.

Ok, I understand your point better now.  So if we want to evaluate log
at -1 - 0.0*%i, we would actually write

    log(-1+(0/-1)*%i),numer;

or something like that?  But we might not get the answer we want if it
were written log(-1.0-0.0*%i)?

I think your proposal works for symbolic manipulation.  I was
concerned about floating-point evaluation and how the user would
interact with that with floating-point numbers.

Ray