In railroading the word "station" has many meanings.  The meaning depends who you are talking with:  

   - In 1972, when I was teaching (railroad) surveying to first-year-engineering students, and you said "station" I would have first assumed you meant a survey point whose latitude, longitude and elevation were known.  i.e. survey stations along the track.  

   - In 1970, when speaking with a friend who was a conductor on CP, a station was "any named point".  If might be a building, or it might be just a sign (no building in sight).  Or it might be an entire geographic area ("Toronto"), within which there were many station-buildings ... and zones.

   - In 1974, when I was working on an automated dispatching system, things got a lot more complicated.

In ProTrak:

A)  Traffic:  

If you are modeling just the traffic on your railroad, all you need to do, for a station is enter a) the name and b) a zonecode for that station.   The station is a "geographic area", with one zone code.

The zone code is needed if there are customers, staging tracks, junctions or interchanges at that station.  The zonecode is what "stitches" your railroad together.

B)  If you are also modeling dispatching activities on your railroad you plainly need "distances".  There are two "distances":  1) the length of a passing siding and the associated parallel main track , and 2) the distance between adjacent passing sidings.  The total of all these "distances" is the length of your railroad.

What to call these two distances?   And what to associate these distances with?

For an answer we can go to computer-based train performance calculators, traffic programs, or automated dispatching systems.  There we find, that for ease of memory and association, the "main-track/passing siding" distance is associated with a "station" (irrespective of where the building is.  Or sign.)   

Now, for the distance between passing-sidings... a)  the distance between main-track/passing-siding has >no< name by definition.  Or it would be a station in its own right.  No name.  Therefore this distance is associated with some named location... which is the main-track/passing-sidng-distance.    b) there are two such distances, one to the east and one to the west.  So we pick a definition, somewhat arbitrarily, noting that "1" is (and odd numbers are) for an "east to west" direction: the distance >to the east< between passing-sidings we associate with the main-track/passing-siding.    We need a label for this distance (shorter than "the distance between main-track/passing-siding"). Let's call this distance the "interval".

Where is the separation point between "main-track/passing" and the "interval".  Turning to the operating rules we see the concept of "clearing up" "at a switch".  We pick the clearance point at the end of the "main-track/passing-siding" as the separation point.

So know we are at the situation as shown on the defintion diagram on page 42 of the Recipe Book.

There is some additional utility for this method of schematizing the "main" (dispatcher-controlled) trackage defined this way.  Double-track, or two main tracks (or three or four main tracks).  If we set the "interval" to zero, then what we have is a series of "two parallel tracks" (or three or four...)

Note that this is only about "dispatcher-controlled" track.

~~~~

There is a lot more, about "station" definitions.   But are we OK so far?

             Jim