Do you want 

What is (1 + 5 i) + (2 - 3i) =

Answer   abs =         arg = 

where you want the student to enter the answer to abs and arg in either numerical form to # sig figs or in terms of sqrts and atan.

Also the complex numbers we are adding can be randomised.

Is this what you are looking for?

@Doug Meade Thanks for your feedback. I'll contact you when I post an article that covers these issues.

@Doug Meade I have come across different approaches to this sort of problem. Some lecturers will allow their students a number of syntax errors on their first attempts to use Maple T.A. and mark their answers correct by overriding the grade in the Maple T.A. gradebook. This of course increases the workload for the teacher but does give the student a better experience. They may allow students 2 syntax errors in week 1, 1 in week 2 and then none continuing forward with the expectation that the students will pick up the syntax with time and familiarity.

An alternative approach, which requires some front-end work by the lecturer, is to construct the question to be more forgiving to syntax errors. I think at first this may give a better student experience but may have the downside that all lectures at an institution may need to follow the same pattern, otherwise what works in one course would not in another.

Do you have any thoughts on these approaches?

For a future post I would be interested in taking the list of examples forms you gave and demonstrating how to make this question more forgiving on syntax. Would that be something you would be interested in?

@Doug Meade Thanks for your comments. 

but how does Maple TA respond when the student's response is not an initial value problem in correct Maple syntax?

Can you explain this statement further? Do you mean what happens if the student gets the answer wrong?

For the grading code. I agree this can be an interesting issue. Probably the most robust grading code I have found is 

ta := dsolve({$TA});
rep := dsolve({$RESPONSE});
tasimp := solve(convert(simplify(expand(simplify(ta, symbolic)), symbolic), expln));
repsimp := solve(convert(simplify(expand(simplify(rep, symbolic)), symbolic), expln));
evalb(tasimp = repsimp)

This would catch cases like this

$TA      ->   diff(u(y), y, y) = 0, u(0) = 0, u(h) = U*cos(t)
$RESPONSE ->   diff(u(y), y, y) = 0, u(0) = 0, u(h) = (1/2)*U*(exp(I*t)+exp(-I*t))

Which are not solutions to the above flow but I have just included as an illustration.

@Just Sorry I should have been clearer about the import process. It can be confusing. To import the question you need to click Assignments then click the Import Content link on the left hand side.

You then click Import and choose the zip file that I provided. This will present you with a screen detailing the content in the zip file. Selected everything and proceed.

That was just for future reference. It seems you have sucessfully created the question manually.

You need to be careful what sort of syntax you accept in this box. You need to choose Maple-syntax and text-entry only. This is an option for the answer box itself. Depending on how you created the question, there are two places this opton could be.

Fig: Changing expression type for a question designer response area

Fig: Changing the expression type for a Maple-graded question type.

By default this is set to formula so does not expect explicit * symbols for multiplication. The effect of this if Maple T.A. attempts to interpret what a student enters, so

a5 = a*5 = 5*a

5a = 5*a

which are, in the end, mathematically identical.

Fig: Expression type: Formula

For what you require this is not the desired behaviour. You can change the algorithm to 

$TA="y2=x1/x2";

Now as long as you change the accepted syntax on the answer box, this question should work as you required.

Let me know if you have any problems. This can be a quite tricky issue and I agree that you should hide your students from complicated syntax where possible.

@1246 

The only solution I could find is

$ep="ε⁺";
$sigc="σ^*";
$sigp="σ⁺";
$eps="ε^'";

Can you post the code you used to use to generate sigma along with the results and the code you now have to use and the results this generates?