The article is certainly correct. Of course, in order to be able to make all of those choices it is mandatory to understand the application. MY best example comes from years ago, which was selecting rear wheel bearings for a custome made motorcycle rear wheel. IT turns out that machining the wheel is not really that hard, but in order to pick the right bearing I had to understand the loading, in addition to the speed and chain tension. At the time I had not been to engineering school yet, besides that, it was not in the realm of what they taught EE students. Ultimately I picked bearings intended for the front wheels of a small car. This was a good choice because they lasted and never gave any problems. In addition, if they had failed I could have purchased replacements in almost any town in the US.
Were they "overkill"? Possibly they were more than I really needed, but isn't reliability worth a lot?
Great summary of the constraints of design envelope, load, alignment, stiffness, and precision. Keeping an eye on these issues will reduce early failures, which is the plague of all bearing applications. And, of course, don't forget lubrication.