Which PageRank formula should I use?

I see a lot of people using PR(A) = 1 - d + d * \sigma PR(E)/L(E) as the pagerank formula. But it's really PR(A) = (1 - d)/N + \sigma PR(E)/L(E). I know there won't be that much a difference since if PR(A) > PR(B) then it's still the same whichever formula you use. Also in Larry Page's paper on PageRank he said that when added together all pagerank开发者_如何转开发s should equal 1

If you want that all values sum up to one you should use the (1-d)/n + d * sum(...) version.