You'll need to post the complete code so that we can duplicate your problem. You can post a worksheet by using the green uparrow tool, which is the last item on the second row of the toolbar in the MaplePrimes editor.

@acer 

Okay, I made the tables explicitly constructed as tables, which takes care of all your objections except the last. Regarding your last objection, I find it rather distasteful for code to use more variable names than are necessary. For one thing, it interferes with the garbage collection of the tables. A better solution is to not use names that reveal the underlying data structure.

@tomleslie Yes, your worksheet works in Maple 16 and gives a very reasonable answer. When N is increased to 100, it gets the answer to 4 decimal places.

I don't understand the significance of the ScatterPlot3D and the lowess surface.

kegj: If you want me to check your code, you'll have to upload the most recent version. The version that you most recently uploaded obviously doesn't work because you've mixed up lowercase and uppercase letters.

This extra space is a long-standing and oft-reported bug. AFAIK, there is no known workaround.

@kegj In Maple, the familiar constant Pi is spelled with a capital P; when you spell it pi, it just becomes a regular variable name.

@tomleslie The different behavior of Pi in floating-point expressions is intentional and documented.

@maple fan You are trying to apply the paradigm of the plot command to implicitplot. There is no algorithm by which you can apply adaptive plotting to an implicit plot.

The numpoints option applies to the overall grid, not to the individual curves. One thousand points means that a 33x33 grid is overlaid on your x and y ranges, and for each cell, a check is made for each curve. At numpoints= 10000, that grid is increased to 101x101. Option gridrefine causes each cell to be further refined if it is suspected of containing a point. Please read the help page ?implicitplot. Be sure to pull down and read the Options section.

Note that your red curve is highly complex. It is quite diffiuclt to find real-valued points along the curve. Try doing it.

You probably should be using fsolve instead of solve. You can supply a guess with fsolve.

Please post your code in plaintext form or upload a worksheet.

@Axel Vogt Using 500 digits and floating-point determinant, I got the same answer.

@Markiyan Hirnyk 

You can assume that if the code is presented here in plaintext form that it's meant to be executed as 1D input.

@Arno wrote:

The fsolve command only returns 1 solution, so what I'm trying to do with the loop is to determine the first nLimit solutions > 0

There is an alternative to fsolve for exactly this situation. It's called RootFinding:-NextZero.

@nm To my reading, the OP is only interested in the unit interval, and the function is real-valued there.

@mzaman 

Let's start with the 1D input. After you've completed that we can discuss evalhf.

Go to the menu Tools -> Options -> Display -> Input Display, Select "Maple Notation". Click on the Interface tab. Go to "Default format for new worksheets". Select "Worksheet". Click on Apply Globally at the botton of the page.

Now use the 1D input worksheet that I already posted on your previous Question. Edit it to reflect the new parts of your program. When you finish that, repost it, and I'll put in the evalhf.

I can't access your paper from that website. It won't let me join as myself---it requires an institution. So, would you please provide an accessible version of your paper?

@Carl Love 

Since you claimed in your Question that e and d were inverses, I thought that you would have some response to my claim that they aren't. How did you generate e, given p, q, and d?