Re: proposal to change acosh



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

    Richard> Wouldn't it be sufficient to do  (eql 1 (cadr x)) instead of
    Richard> (and(fixnump...)(= 1 ..))  ?
    Richard> If (cadr x) were not a fixnum, it certainly wouldn't be eql to 1.

Yes, this is much better.

    Richard> ..  (Barton's correction??)
    Richard> .. also, I think it would be appropriate to extensively comment this
    Richard> function, while
    Richard> ... you have figured out what it does.

You mean testt?  I didn't actually figure it out.  I just traced
simplifya and friends until I found the point where 1.0*x was getting
converted to x, which happened in testt, and noticed there was a onep1
that was checking for 1, 1.0, and 1b0, so changing to check just for 1
was obvious.

I also didn't really figure out how csc\Sec-plus and csc\Sec-times
work.  I just traced the functions and sin\Cos-plus and sin\Cos-times
to see what was happening, and noticed that if I reversed the args, we
would get the right results.

But yes, everyone making a change to a function really ought to
comment it, if you've only figured out a part of what it does.

I have been rather lazy about that.  I will try to do a better job. 

Ray