[Talk-us] Rail westerly

[stevea]

A not-so-brief review of what has happened in the ten days since I 
brought this up.

I have had fruitful conversations both on- and off-list with several 
responders to my "shout out" about improving rail in California (in 
OSM).  Largely speaking, this resulted in better harmonization about 
what and how to tag existing (usually from TIGER) rail lines so they 
better render with the rich/vivid renderer ORM.  I might also 
continue that conversation here in talk-us, glad as those who 
answered here are followed by more (regionally around the USA) as 
they might wish to further contribute to better USA rail.

A "larger gestalt" approach seems next, one mightily contributed to 
with many broad voices.  Regionally diverse, containing an 
intersection of "rail fans" and "rail professionals" and "active in 
their area mappers" as a good start.  A new OSM WikiProject, anyone?

We are getting better at identifying what to put into the name= 
owner= and operator= tags, as rail ownership and usage (sometimes via 
leasing agreements, public ownership/private usage or vice versa and 
complex trackage usage arrangements) can be opaque and difficult to 
determine.  This still seems like an endeavor of "do our best, even 
as it results in needing continuing conversation."  California's 
Public Utilities Commission crossing data spreadsheet (>13,500 rail 
crossings in our state) is a helpful beginning to this.  However, it 
contains old/obsolete data and refers to disused/abandoned tracks as 
if they have "live" rail operations.  Much can be done with the 
tedious tracing of an entire line along its length via its road/ped 
crossings, making  this sort of work possible, but fraught with the 
peril of errors and tedium.

It remains subjective interpretation as to what constitutes 
usage=main and/or usage=branch on some rail segments.  It may or may 
not be the case that major (e.g. Amtrak) passenger rail always gets 
usage=main, or even usage=branch.  Usually it is one or the other, 
but "not always."  Also, there are segments with "Main Line" in their 
name which have light industrial (or even no) rail traffic where it 
seems the semantics of "usage=main" does not readily apply.  This may 
result in a usage=branch (or even usage=industrial) tag on a line 
with name="XYZ Railroad Mainline."

The ORM tagging wiki suggests that a hierarchy of relations be 
applied:  infrastructure, railway route and train route.  In the USA, 
our TIGER data entered rail infrastructure as tags on ways rather 
than as ways collected together as members in an "infrastructure 
relation."  While this differs from the suggested ORM tagging, I 
believe it is OK to continue with this methodology in the USA:  rail 
tags on ways, railway routes and train routes in route relations. 
ORM renders with this slightly different tagging just fine, as do 
most renderers with a similar structure (e.g. the way the USA tags 
highways with way tags and routes with relations).

When I started this ten days ago, California rail in OSM via ORM 
looked highly incomplete to me:  perhaps what might roughly be called 
"35% done."  Now, San Diego, the Bay Area (especially) from about 
Santa Rosa to Salinas and (to a lesser extent) greater Los Angeles 
look much, much better.  The Central Valley continues to improve, and 
far northern California is still in need of attention, but I would 
say California rail in OSM is now something like 60% done.  That's a 
fair improvement for ten days of "occasional hobbyist effort during a 
holiday period" and encourages me that other volunteers who wish to 
do a similar improvement of their area's rail might get a lot done in 
a relatively short period of time.

I may continue to tap away at "improving California rail," but it is 
a large mountain to climb.  Obviously, for the whole of the fifty 
states, it is a very, very large mountain.  One step at a time:  be 
encouraged, not discouraged!

Truly, if anybody is looking for a "meaty" project (which might very 
naturally become a new WikiProject), please consider putting some 
effort into improving rail in the USA along these lines.  Yes, ORM 
rendering only daily and the vast number of track miles makes this 
slow going, but IMHO, results are well worth the effort.

SteveA
California