If you are going to zero the bore by indicating off of 2 points down the bore, the distance between those 2 points will be directly slaved to your accuracy of index. For example, if you have a curved bore (almost all are just the extent is variable) and you indicated 1 inch into the bore and then travel down only 0.250" further to check alignment again, you will have very little run-out to detect even if the bore is in fact off axis to the spindle. Accordingly, the further apart you can stretch your 2 check points, the more easily and accurate you can detect any existing curve and dial it out. ....... so the longer your indicator can reach, the longer you can stretch those 2 points.
If you subscribe to that theory, trying to detect and dial out any curvature with a short stem indicator is almost as bad as indicating off of only one point in the bore and assuming that the bore immediately downstream from that point is running with no axial run-out the spindle.
Truth is, that most better barrels have small enough curvature that you could likely get away with the lesser of the above and still never see any accuracy issues in the real world considering all of the other variables.
That being said, we should all try to execute our work with best practice and best technical theory we can to insure concentricity on the finished product.
This is just my own personal opinion on the way I should setup and cut. I have definitely found that a lot of this is like religion.. . . . . we can all have variations of how we believe things are and still live our lives happily with the results we see.