[Talk-us] Rail westerly

stevea steveaOSM at softworkers.com
Mon Dec 29 22:27:29 UTC 2014


Hi Charlotte (and talk-us, and apologies for the length of this):

>	Yes, any explanation of the difference between "main" and 
>"branch" lines and "owner" is appreciated. I find myself rather 
>confused at the moment.

It is complicated.  The owner= tag is for the actual landowner of the 
railroad.  This usually includes land (maybe twenty to a hundred or 
so feet on either side of the rails, sometimes -- but certainly not 
always -- a polygon tagged landuse=railway), tracks, ballast, ties 
and signalling equipment.  Due to the real complexities of ownership 
(which may be private, like a railway corporation, e.g. Union 
Pacific, BNSF -- there are hundreds -- or public, like Port of 
Oakland, Santa Cruz County Regional Transportation Commission, 
hundreds of others), owner= can be hard to determine, but there is 
always only one owner.  Adding to the complexity, there are also 
leasing agreements the owner may arrange with other entities (again, 
potentially both public and private) like railway companies, public 
agencies, and so on.  These leasing / trackage rights arrangements 
are where it can make sense to add an operator= tag to a rail way 
with an owner= tag, if they are different.

The name= tag often includes the "subdivision" name of the "line" of 
the rail.  There is a hierarchy, but it is loose:  at the top is a 
rail company (again, like Union Pacific or BNSF).  These "owners" 
(but could also be "operators") often are referred to as a "Primary 
Rail Organization" as a top-level entity.  There are various 
abbreviations for these, like UP and BNSF, or even TTBART for Bay 
Area Rapid Transit, and I'm not sure what (federal?) standardization 
there is for all of these abbreviations, but they do seem to be 
largely consistent across much of the documentation I see.  It seems 
like it might make sense to put these abbreviations into a ref= tag, 
but while I haven't seen that widely done, it is beginning to happen 
(perhaps for "shield tagging" like we see with MapQuest Open?)

OK, so below owner/operator/ref (the Primary Rail Org), there is what 
is known as "Division" or "Line" name.  And below that, there is 
"Subdivision," especially among the very large/major railroads like 
UP and BNSF:  these "majors" almost always have both Divisions and 
Subdivisions.  However, medium-sized and smaller railroads might go 
straight to "Subdivision" as the next level down from their top-level 
ownership -- they might even go straight to something they call 
"Lead" or "Line."  Below Subdivision is "Lead or Line Name."  These 
are usually spurs and small industrial segments that branch off of 
usage=main or usage=branch segments.  It is these names (except the 
Primary Rail Org, which goes into the owner= and/or operator= tag), 
especially Subdivision, is what should be put into the name= tag. 
This holds true for infrastructure (ways in OSM with a railway=rail 
tag) which is also tagged usage=main AND usage=branch, which I'll get 
to below.  Now,  it is also true that if an industrial lead (like 
"Spreckles Industrial Lead") is actually named that, EVEN IF it is a 
branch/spur off of a usage=main or usage=branch (its usage tag should 
be usage=industrial or service=spur) you should put the name of it 
into the name tag.  So in the above example, the spur off of the 
Union Pacific's "main line" through Salinas (with tags name=Coast 
Subdivision and usage=main) has name=Spreckles Industrial Lead.  I 
know, it is confusing, but I think you can get the hang of it.

Just enter owner= as the Primary Rail Org (unless there is a leasing 
agreement, in which case you might also use an operator= tag), and 
put into the name= tag the "name" of the "line" which is usually a 
Subdivision or Lead/Line name.

Now, adding a usage tag (usually usage=main, usage=branch or 
usage=industrial) is even more tricky:  it isn't really clear what is 
usage=main or branch, especially in the USA.  Often, usage=industrial 
is more straightforward, especially if it is clearly a small (often 
urban) spur off of a main or branch line that disappears into or 
along the edge of buildings in an industrial zone.  So, a place to 
start thinking about where to add usage=main are the major "lines" of 
the major railroads which connect larger geographic areas:  not 
county-to-county, but more like state-to-state or region-to-region. 
We (in the USA) are just going to have to "figure this out" as we go, 
as it is a bit of subjective interpretation where to best add the 
usage=main tag.  At a hierarchical level distinctly below usage=main 
is usage=branch.  Sometimes, it will be clear after sketching out a 
major line being usage=main that there is a "branching structure" 
that has rail connecting or radiating outward from the usage=main 
tracks, and it may very well be that adding usage=branch to these 
segments is correct.  (Alexander Jones' work in the Central Valley 
illustrates this).  But again, we are going to have to "feel this 
out" as we go along.  This is where we need "rail fans" or "rail 
professionals" to correct us where we are wrong, as the structure of 
the network is what we are defining with these tags (main and 
branch), but people who really KNOW rail have a real sense of what 
"the" (or "a") more correct instance of this likely is.  We just have 
to listen.  Good stabs at usage=main and usage=branch can be 
carefully undertaken by us as an initial first step, and I think we 
should do this, but we must be willing to change the structure when 
told by those more knowledgeable that we got it wrong.

Finally (in this missive!) there are service= tags:  service=spur is 
essentially equivalent to usage=industrial, service=siding is for a 
track (often in a rail yard or at a station) where "through traffic" 
can pass another "parked" train, service=crossover is for a short 
segment of track that "switches" from one rail to another parallel 
rail and service=yard is for tracks that make up a rail yard.  The 
service= tags are "more rare" compared to usage= and name= tags 
(essentially, all rail ways should have usage= and name= tags), but 
when you need a service= tag, you need it, so tag it!

>	For instance, why does the BNSF line across Arizona and into 
>New Mexico change to "Gallup Subdivision" at one point to the west 
>of the city of Gallup? It seems to be still the BNSF main line. 
>Further, other "subdivisions," such as the Springerville (Ariz.) 
>Subdivision, clearly seem to be branches. So, when is a line a 
>subdivision?
>	If I get a handle on this and other distinctions, I can make 
>corrections (or explanations) ... I think.

To be clear:  something can be named "XYZ Subdivision" and STILL or 
ALSO be usage=main (or usage=branch, for that matter).  The word 
"subdivision" is what a lot of people (I'm discovering) seem to 
prefer as the "name" of a segment of rail -- whether it is a "main 
line" (usage=main) or a "branch line" (usage=branch).  In short: 
"subdivision" is a convenient name, without regard to where it might 
be in a usage= hierarchy.  It is the usage= tag which defines 
hierarchy in OSM's rail network, not how it is named (even if that 
name has the word "division" or "subdivision" in it).  Got it?  Whew!

Just re-read this a couple times (if you must) and with these 
guidelines in mind, try a few fixes if you think the map needs them. 
ORM rendering only once a day is painfully slow from an interactivity 
perspective, but just have patience and you'll get the hang of it 
after a few edits / days of rendering.

If I had to answer Charlotte's specific question, I'd say that the 
Gallup Subdivision might have a usage=main tag, but Springerville 
Subdivision might have a usage=branch tag.  Both are named 
"Subdivision" but that's just how the rail company (owner=, 
operator=) puts them into THEIR hierarchy.  OSM's hierarchy that is 
expressed with a usage= tag is what we're doing our best to tune up 
here.  We can get it a little wrong, and we should correct it when 
appropriate, but we can sketch in what we think are main lines and 
branch lines for now, if we keep all this in mind:  add owner= (and 
operator=) tags and especially name= tags first, usage tags come 
after this and a bit of thought and/or "more network knowledge."

So, to Mike Henson about Oklahoma:  one line might be named one, a 
few or even several "subdivisions," but it is the usage= tags we are 
concerned about which outline (in ORM) the "network structure" of 
rail in the USA.  These might all be one single owner= (or they might 
not be), but if they are all a "logical main line" then they should 
all get "usage=main."  Again, these are a bit subjective (right now), 
so let's just do our best to keep all this in mind.  We can 
mid-course correct if/as necessary.

>         I'll try to help with the wiki, particularly from an English 
>standpoint, my specialty.

Excellent!  The USBRS WikiProject was my "first" and I had help from 
Minh, who created the pages out of nothingness (which you, Charlotte, 
might do for WikiProject USA Rail):  Minh planted the seed, and I, he 
and others have grown it into a thriving and mighty tree.  I do hope 
you can plant a similar seed and we can all grow a USA Rail wiki!

>         I'll also try to pick up more knowledge to contribute to the 
>wiki, though I'm a bit busy with the HOT web pages at the moment. We 
>don't have to hurry--there's no crisis--so probably it's better to 
>take  our time and get as much as possible from rail experts. I'm 
>one of the "area" mappers, but I know little about railroads.

Correct:  this isn't burning a hole in urgency (like some things, 
especially HOT, which IS truly hot!)  I'm no rail expert, either, but 
it is becoming an interest of mine as we begin a potential passenger 
rail project locally here in my county of Santa Cruz.  Good maps and 
map data can only help further, future endeavors.  I often say that 
the USA has a 20th or even 19th century rail system, and the rest of 
the world is clearly in the 21st.  We don't have to spend billions or 
even trillions "catching up" (well, yes, we do...) but we can 
certainly display "where we are now" with good geographic rail data 
and nice renderings like ORM show us -- without spending anything 
more than good volunteer effort!

Thank you for your enthusiasm!  If my explanations aren't clear, or 
you have further questions, please ask and I'll do my best.  But do 
take a look at that CPUC spreadsheet (it is HUGE!) as it really can 
be helpful (though tedious), and I can send you a very early version 
of some of the "branching spreadsheets" that Alexander and I are 
working on for California (as you are in LA).  We know these probably 
are not 100% correct, but we'll both take 75% or 85% (where I think 
they are now) and start putting those into OSM and see how far that 
takes us.  Actually, Alexander writes me he is becoming a full-time 
student again in about two weeks, so his work in Central Valley will 
rapidly fall off to zero very soon.  As for me, I really do have 
other things to do, so while I might tap-tap-tap at California Rail 
in OSM a bit, I can't give it my full attention.

Cheers,

SteveA
California



More information about the Talk-us mailing list