@acer Thank you for clarifying that for me.

@ecterrab Thanks for repeating; I missed that. Muito obrigado!

@ecterrab How do I set one of the constants = 7 for example?

I tried c[1]:=7 but that does not work.

I corresponded with tech support about the 4/6 post about the length command and they clarified it for me.

"Maple development team say that the "formula" used for length was not designed to be clear to users, but instead, to be as fast as possible to compute given the internal representation. It is used a lot for intermediate steps in internal computations."

In other words, it is mysterious-not meant for casual users.

I do not understand the example:

length (x+2y) = 9

but I note that
length(x+1) = 9 also.

(Maple 2022)

In fact, it seems very mysterious.

@Joe Riel Thank you for explainng that.

@mmcdara

In fact I would be happy with [123,0,321]; they are data points in range 0 to 1000.

I am still curious why convert("531",integer) gives that error message.

Origin of problem: I generated a long list of 3 digit numbers that I wanted to plot. However, I did not notice that there were no spaces or commas betwen the entires. I was trying to salvage all that typing.

I tried these simple procedures that fail and I was curious about Maple's pattern matching capabilities. (I think this would be very simple in Mathematica.)

f:=proc (x) if x= "y" then y end if end proc;
or

f:=proc (x:: "y")  y  end proc;
 

@dharr 

Since h(0,x) =diff(f(u,x),x) at  u = 0 and f(0,x) = g then h(0,x) must be diff(g,x) to be consistent. I should have mentioned that I tried that for another dead end. Here is the message:

Error, (in pdsolve/numeric/match_PDEs_BCs) cannot handle systems with multiple PDE describing the time dependence of the same dependent variable, or having no time dependence

@Preben Alsholm Hi,

Thankk you for your response.

I checked kernelopts(floatPi); in both Maple 2019 and 2020 and got true in both cases.

However, I realized from your answer that I got the wrong answer when Digits was set to 10 but if I put Digits:=15 all was well. I guess I was seeing round off error.

@Carl Love  That is quite subtle. Good to know. Thank you. I guess the longer displayprecision

is the command I will have to use.

@acer 

Thank you for explaining my error. As Tom Leslie also pointed out using evalf( ....,5) was my problem.

From the syntax-it sure looks like a composition- I assumed that since  f(g(x)) is computed by first computing g(x) then applying f to that result evalf( eval([W,Wp],nans(1.5)),5) would be computed by first evaluating eval([W,Wp],nans(1.5)), then applying evalf..., 5) to that answer.

In my work I was using dsolve with Digits:=35 and error tolerances of 10 ^(-14). I guess evalf overrides those instructions somehow. So be it.

The interface setting does resolve my issue.

@tomleslie 

I do not see any major changes you have made. Could you please explain? I also do not understand what you dislike about my code or where it might go wrong. f'^2 cos(x) /f^2 > 0 for any any f if 0 < x < πi/2.

@Carl Love evalf_error.pdf

does this help?

evalf_error_3.mw