I find that the best way to solve that is to make the strategy define its desired state, and then you have another component that diffs the current and desired states to identify the necessary transactions to send out as well as their priority based on the nature of the diff, then finally go to a scheduler.
The point of the article is to show building a calculator requires a CAS, which should have been obvious to anyone with a basic understanding of how a calculator works.
The premise of the article is itself somewhat bogus, but I suppose there are programmers today who never had to work with a graphing calculator.
While RRA is an interesting approach, ultimately it wasn't sufficient.
Re-using an off-the-shelf CAS would have been the more practical solution, avoiding all the extra R&D on a novel number representation that wasn't quite sufficient to do the job.
reply