🚀 go-pugleaf

On Fri, 18 Nov 1994, Sid Birnbaum wrote:

> Subject: Re: Graph-Ti;  y = x^(1/log x)
>
> Concerning earlier posts by Fred Worth and Russel Blyth on this subject:
> I can't explain the anamoly, but I can certainly add to the confusion.
>
> One would think that in doing the calculation for the graph, the TI-85
> actually calculates
> 		y = exp(ln x/ log x)
> However, graphing this function (using the same range settings) gives
> y = 10  for all calculated values.
>
> Further, the graph for y = x^(1/ln x) gives no surprises : y = e, for
> all calculated values.
>
> As a side issue, calculating the subject  function for x = .6595744681 ,
> which is 1/10 of a troublesome value, gives 10.
>
>

 It seems to me there is a bug in the coding of at least my TI-85 when
evaluating the above expression.    First I have a 8.0 ROM.  You find the
ROM version of a TI-85 by going to the MODE menu then  [alpha] S
  I reccomend you push the [on] button when it asks about the self test.

  Now the case in point.  When the TI-85 graphs a function it calculates
the difference between xMin and xMax then divides that into a bunch of
parts.  It finds f(x) at part one then continues.  Because there is a bug
in the evauation of y1=x^(1/log x ) some values give the wrong answer.
Most values give 10, but some don't.  To demonstrate enter y1 on the graph
menu as x^(1/log x)  then set the following in range.
  xMin = 1     (most important for this example)
 xMax = 185    (most important for this example)
 xScl = 10     (any value here)
 yMin = 9
 yMax = 11       puts the ten line right dowm the middle
yscl = .1      (or whatever)

 Using Graph with above values on my TI-85 show four of the "bug
values"
 In order for the bug to show the TI-85 has to be right on the value.  If
it is just to the side of it then you get the correct answer.  Again
refering to my version 8.0 rom TI-85 change  xMax = 185.1 then regraph.
I know of no other function where you increase the RANGE 0.05% and the
graph changes so much.
    Other ranges show other bug values.  I played for a half hour before I
stumbled on the above FOUR bug range.  Lets have a contest!  Who can find
a FIVE or more bug set of range values?
Elwood Gruschow
gruschow@ncsa.uiuc.edu