You are missing the point. All instances of a name are important, most of my relatives have different AKA names based on where they lived at a given time. I always add one as the primary name, normally this is either their birth name or the name that most people knew them by while they lived. Then add the other names with a "type" of AKA or married.
Several of my relative, including myself, have "pen", "stage", or pseudonym which I also record. All of these names are still recorded as a name, but with different "type" values.
As far as spelling changes I record these as names as well not in the AKA field but as a NAME with an attribute of AKA. Using the same data construct is good database practice and can be better used in reports. Because then it is either a list of all name with their type attribute or the first ( primary) name when you only want one name listed.
So again... It's not that I want to eliminate AKA I want to organize it in the correct construct which is name.