User blog comment:Googleaarex/Tetrational Arrays/@comment-25418284-20130311001458/@comment-1605058-20130312153540

It's pretty easy to define prime blocks of ordinal a: for prime p take first p ordinals in fundamental sequence of a. If there are any limit ordinals on this finite list, take first p elements of their fundamental sequences and so on. For every succesor on final list its predecessor is guaranteed to be on that list too. Do you think this may be good approach?