I think so too! Coming from cvx/YALMIP, I find it hard to transfer to JuMP. I am documenting these, especially for folks coming from control/estimation and are used to cvx/YALMIP. Perhaps it can be included with JuMP examples in the future.
With Dualization, things are better. However I am still running into difficulties in implementing constraints such as s\ge0 where s\in\mathcal{R}^n.
The documentation for Dualization is probably aimed at developers at this time.
I have to spend some time on this to figure these basic things.
It may be better to dualize the problem just before calling \texttt{optimize!(...)}. But then I don’t know how to recover the primal variables from the dual solution.
Where can I get information about this?
I am going to mark this topic as resolved.