The article seems geared towards someone working from a book or guide at their leisure. Building things up from first principles is a great way to learn but I personally never have time for the approach on the job.
I think what the article does get right is a fundamental technique to solve a problem when you're stuck: reduce to the essentials, get your terms / naming things right, write a teeny self-contained thing (I call it 'building a model') to try it out. When the solution is understood and has proven to work, integrate. Especially when you're hard pressed for time, this may prove faster overall than the approach where you fumble with many screws.