Recently I asked this question and quickly got back some excellent responses. I asked the question because I came across a paper by Eric Rowland called "A Natural Prime-Generating Recurrence". The main results on page 4. It appears there was much fanfare in academic news when the paper was released which turns out to be basically the question I asked yesterday except, instead of subtracting the lpf, he added it causing his function to climb indefinitely, whereas mine is decreasing.
So basically Rowland's iterations is $$f(n) = n + lpf – 1$$ starting at 5 (where lpf = least prime factor of $n$) which yields: $$5→9→11→21→23→45→47→93→95→99→101→201→203→209…$$
The sequence above is exactly the sequence Roweland has in column 2 on page 4 of his article without the duplicates.
Am I missing why Roweland's sequence is a significant discovery?