Aliases and equations
= defines an alias, == writes an equation
In a @def block, a line name = expression defines an alias: later lines can use name in place of the expression.
@def begin
t ∈ [0, 1], time
x = (q, v) ∈ R², state
u ∈ R, control
q̇ = v(t) # alias: q̇ stands for v(t)
v̇ = u(t)
ẋ(t) == [q̇, v̇] # the dynamics is an equation: ==
...
endDynamics, constraints and boundary conditions are written with == (or ≤, ≥). An alias has a name and no arguments.
The common slip
Writing the dynamics with = instead of == is easy to do:
ẋ(t) = [q̇, v̇] # ParsingErrorJulia reads ẋ(t) = … as the short form of a function definition, which has no meaning in the DSL. The parser refuses any statement of the form f(args) = body with a ParsingError that shows the line as it was written and suggests the same line with ==:
ParsingError
│ Line 7: ẋ(t) = [q̇, v̇]
│ `=` defines an alias, which is a name without arguments; an equation or a constraint is written with `==`
│
│ Hint write `ẋ(t) == [q̇, v̇]`
└─This holds for the derivative of the whole state (ẋ(t), ∂(x)(t), derivative(x)(t)), for a component (ẋ₁(t)), and for any other equation such as a boundary condition x(0) = 0.
An alias is defined once
An alias cannot be redefined: a second c = … after c = … is refused with a ParsingError that names the alias, even when the value is the same.
ParsingError
│ Line 5: c = 7
│ the alias `c` is already defined
│
│ Hint rename the second alias, or remove the first one
└─Reading the line in an error
An error names the line of the block it comes from, Line n: …, printed as the statement was written, without the internal comments Julia adds when it lowers a function definition. The aliases that @def itself provides (derivative for ∂, in for ∈, <= for ≤, …) are shown in their substituted form.