Just to start the ball rolling with what my cousin calls a "suppository" observation:
Could it be because adding numbers twice is usually easier than subtracting from an arbitrary null point even once?
For example, if you want to reckon Trenton to Wilmington mileage, you might have to add Trenton-to-Philly and Philly-to-Wilmington under the revert-to-zero system. But that's probably easier than using a system under which zero is Penn Station, NYC, and the New York - to - Trenton mileage would have to be subtracted from the (already upwardly accumulating) mileage between Trenton and Wilmington. It's much easier to add than subtract, particularly when not using paper -- and of course impossible to calculate from an arbitrary point somewhere else if you don't know the distance between (null/terminal/origin) and where you start your own count.
Another random thought: could the system that reverts to zero at each (important) station, then starts counting up again, have something to do with calculating a railroad's labor needs? Show mileage elapsed between two stations, double it, and you can tell if a crew turn is possible in terms of permissible distance covered. I guess division points would be important under such a set-up, too.
But I'm just guessing; hopefully you'll hear from the pros and from people from different roads. - al-in-chgo
In Canada on CN and CP Mileposts have NOTHING to do with how far away from a city you are. It is simply flat out the distance you are from the start/end of a Subdivision.
All subdivisions almost always end in a town, but its not always a city. So here you wont see like milepost 1045.5 or 567.4. CN has extended runs, and I think their Kingston sub could be the longest in Canada at 333 miles. IF there is one longer, correct me.
10000 feet and no dynamics? Today is going to be a good day ...
Subdivisions are one thing.And a main line milepost is another.
Here on the Pocahontas Division the main line mile post all begin with an N. which represents Norfolk Va.,which was the main headquarters for the N&W r.r. .It varies around Nagatuck WVa to Kenova,WVa account the main line moved in the early 1900's ,and therefore the initial changed to NA. which stands for Nagatuck WVa., where the "new" mainline started from the east going west from Norfolk Va. The N. was used on the mainline all the way to Columbus Ohio where the original N&W ended.
The branchline out of Kenova ,named Wayne Branch uses the old maine line,and still uses the N. for the milepost until it goes off to the Colemont spur track (Colemont Mining Co.) which is the only indsustry served on the entire 32 mile long branch.I can't remember right off what they use on that branch a C. or a W. it's been too many years since I've been qualifed up that way.
I would think most railroads use the same principal on their mainline mileposts.
Collin ,operator of the " Eastern Kentucky & Ohio R.R."
The NYC&HR had their mileposts NY to Buffalo, then the LS&MS had theirs starting at Buffalo. The PRR had theirs starting at Philadelphia, then the PFtW&C had their own starting at Pittsburgh (and the PT&T had theirs starting at NY Penn, and the NJRR had theirs starting at downtown NY (Jersey City was MP 1)). That's the usual reason for setting back to zero-- you're changing to a different predecessor RR.
On CN there is no distinction between main line or branchlines when it comes to mileposts. EAch subdivision starts over at mile 1 to whatever the length on the sub is. Branch lines are their own sub so they have their own mileposts ... the exact same type that is on a mainline.
They do things differently here in Canada.