🚀 go-pugleaf

Rick Seiden writes:

> I have found an error on the TI-81.  If you set the mode to radians,
and take
> the tangent of pi, you get some really small number (I dont have an
81 in front
> of me, so I don't know what the number is....also, I am not sure if
that what
>produces the error).  I have checked both the 82 and the 85, and did not
>find this error in either.

There is an interesting point here.  You would _expect_ the calculated
result of an expression such as tan(pi) not to be exactly zero.  After
all, there is no way to store (or pass along) an _exact_ representation
of the irrational number pi.  I also do not have a TI-81 in front of me,
but I notice that the TI-82 does seem to evaluate tan(pi) and sin(pi) to
exactly zero.  I assume a more sophisticated algorithm was built into
the 82 and 85.

What is more, the expression sin(1000000*pi) also evaluates to an exact
value of zero!  Very good.  BUT, sin(1000001*pi) yields -4.64102e-8 .
Well, I guess we can stand it....


        ||      RWW Taylor                 rwtnts@rit.edu
        ||      National Technical Institute for the Deaf
        ||      Rochester Institute of Technology
        ||      Rochester NY 14623