restart;
eq:=diff(y(x),x) - y(x)^2 - x*y(x) - x + 1;
r:=dsolve(eq,y(x));
latex(r);

Gives an error "(in fprintf) string expected for string format". The problem seems when the math contains "erf"? Both examples below fail with result that contains this.

Maple 18, on windows 7.

Another example:

restart;
eq:=diff(y(x),x) + a*y(x)*(y(x)-x) - 1;
r:=dsolve(eq,y(x));
latex(r);

Another example, which seems to fail for different reason

restart;
eq:=diff(y(x),x) + a*x*y(x)^3+b*y(x)^2;
r:=dsolve(eq,y(x));
latex(r);

This is better explained with simple example. Suppose I have these 2 equations

a:='a';b:='b';x:='x';
eqs:=[3*a+4*b = 3*x, 5*a+7*b = 7*x];

What I want is to solve for (a/x) and (b/x) from the above, and not for "a" and "b".

Currently, I do this manually as follows. First divide both equations by 'x' to obtain the ratio. I had to do this manually as writing eqs/x gave expression which was hard to work with for what I want to do. I could not simplify it for what I wanted. So I typed this

eqs2:=[3*(a/x)+4*(b/x)=3,5*(a/x)+7*(b/x)=7];   %manually typed !

Now used algsubs to replace (a/x) by new variable z1, and replace (b/x) by new variable z2, then solve for z1,z2 (btw, I had to do algsubs one at a time. (need to look at this later, but not impotant)

algsubs(a/x=z1,eqs2);
algsubs(b/x=z2,%);
solve(%,{z1,z2});

       {z1=-7,z2=6}

Is there a way or function(s) to automate this. The hard part is telling Maple to use the ratio as new variable starting from set of equations.

One annoying feature in using document mode that I found is this. As I hit the return key to execute one command and automatically jump to the command below it (a feature that I like), but after few returns, I find myself at the bottom edge of the window, with the cursor pointing at the last command in the window.

Now when hitting return, the current command is execuated, and the next command below it appears, and it is also at the bottom edge of the window.  The cursor is from now on, always sitting at the bottom edge of the window, and I am not able to fully see the rest of the command and what is below it any more.

What would be better, is to try to keep the window centered, so that as I hit a return, the actual cursor remains in the middle of the window, and the window itself scrolls up. This way the command I am about to evaluate, and anything below it, remain in full view.

What I end up doing now, is when I get to the bottom of the window, is to manually scroll the window down using the mouse, so that now the cursor is up higher, and I can see more the commands which I am about to evaluate, then I repeate this process.

Is there  a way to configure Maple to try to keep the window centered? it is really very annoying to have to keep adjusting the window every few returns, and when one has 100 commands to evaluate, one by one (since I want to see the result before I go to the next command), this process gets tiring.

(does Maple do any usability studies and evaluation on its user interface by any independent outside group?)

restart;
g:=0.88641:
e:=2.53128:
eq:=tan(g)= e*sin(f)/(1+e*cos(f)):
fsolve(eq,f);

gives  1.19749

What command/option do I need to get both solutions like with Mathematica:

g = 0.886461;
e = 2.53128;
eq = Tan[g] == (e Sin[f])/(1 + e Cos[f]);
NSolve[eq, f]

  (*   {{f -> -2.56623}, {f -> 1.19756}} *)

thanks

I find this a little frustrating as I learn Maple, so I think there is a better way to handle this.

I find myself having to keep wrapping expressions with evalf() in order to compare them, since when I use a constant such as Pi in these expressions and then compare them,  Maple complains.

In a large program, one does not know if an expression contains Pi or not beforehand, so is one really supposed to convert every expression to float just in case they might need to compare 2 expressions? 

Let me explain with simple example:

x:=1.2;  #it does not matter if this was 12/10 or 1.2, same error will result.
y:=Pi/3;
if x<y then
   print("x<y");
else
   print("x>=y");
fi;

The above gives the error "Error, cannot determine if this expression is true or false: 1.2<(1/3)*Pi"

So I changed the y assignment above to y:=evalf(Pi/3); or evalf(Pi)/3; and now Maple is happy.

But this for me looks awkward. In Mathematica, I can simply write the same, using symbolic Pi, and it works as is:

x = 1.2;  #even if this is symbolic 12/10 it will also work
y = Pi/3;
If[x < y, Print["x<y"], Print["x>=y"]]

I did not have to write  y=N[Pi/3]  where N[] is the equivalent function to Maple's evalf() which converts its argument to real.

So, now in Maple, I find myself writing evalf() around so many things, since I have to anticipate I might need to compare them in some logic later on and I can't keep track which one has some symbolic constant such as Pi in them or not. While in Mathematica I never had to worry about this.

Is there a way to reduce the need to having to use evalf() so much?
It seems to me, Maple should be able to decide if  1<Pi without me having to write 1<evalf(Pi) ?