The issue with names is that there is still no standard unique identifier of an author (*). I.e. for a paper, the community has more or less converged to use DOIs. However, this has not been the case for authors.
The implication is that there is no way to disambiguate names: is
John Smith in paper A the same
John Smith of paper B?
John D. Smith in paper C and
John Doe Smith in paper D are the same person?
One way to mitigate this problem, from an author perspective, is to choose a name that is unique, or at least unique in your field of research (a real problem for e.g. researchers from China). On the other hand, choosing a name too long can, arguably, be detrimental because readers will not memorize it so easily.
Thus, the recommendation is: use one and only one name throughout your scientific career, and choose that name wisely (e.g. ask feedback from your supervisor). That name doesn't need to be the full name, nor has to be a shortened one; just need to uniquely represent who you are named.
Ideally, I should google for your name and the first occurrence should be either your homepage or your google scholar page (or a paper of yours).
I personally memorize the name easily when it has a first and last name. E.g.
John D. Smith, or
Kendall M. K. Reddy, and only use the middle letters to disambiguate the name, when needed.
(*) there are some attempts, but it is not a standard AFAIK.