Ok, apologies for the “Input Needed” prefix, but since my last rail post - which was clearly uncompelling - had 0 replies , I thought I’d beg this time.
It’s intended to be a fairly exhaustive inventory of what we’ve done so far, and to be the basis for a thorough discussion of where we want to go, for both the Railway and the Historical layers. In other words, if you’ve ever wanted to do x, y, or z related to rails in OHM, then there’s probably a spot for you to comment or review in this doc.
But, enough talk, please check it out & let me know if you have any questions about the doc’s contents or collaborating in HackMD. That’s what I used to create the doc & moving it elsewhere kind of ruined the formatting a bit.
And, please… … add a comment or two here… 0 replies is so demoralizing!
Haven’t had a lot of time for OHM lately, or to think about this specific topic, but my general thoughts on railway styling can be summarised below:
My comments relate to the general OHM view styling and not a railway focussed view. I am not a railway aficionado but are trying to think of rail’s styling in it’s full historical and social context.
Having compared the Americana and OpenRailMap variants I think Americana is a good balance
In OHM rail styling has been generally too dominant with respect to other elements: too visually heavy and with a scale/weight that is insufficiently sensitive to zoom level. Americana is good with respect to weight and colour and the tick marks are sufficient to distinguish that transport network from other networks without visually dominating. The ‘filled-dash’ styling of current OHM also works, but is way too heavy - standard OSM (CartoDB?) looks great.
The discrimination between railway usage is inadequate. Again, Americana is good, I can (just) distinguish between an industrial and main line, but not between a branch and industrial line. Perhaps this distinction could be more clear through subtle weight and grey scale tweaks (or following the OSM styles). Main lines could be a little darker/heavier. I think the adoption of the full-blown colour scheme of OpenRailMap is overkill.
At the largest scale I’d like to able to see the main line rail network of the state I live in as it historically maps the development of that state and is usefully studied/experienced at that scale. For me that’s zoom 7 - see zoom level - ‘state level’ (New South Wales, Australia) - similar size to Texas or Germany, for example. The rail network should have a similar weight (thickness) to the road network at that scale. Given Australia’s sparse population and correspondingly sparse transport network both road and rail could be a little heavier at this scale, but that is not true of other more populated places - Germany, for example - so current weight of road network at zoom 7, with similar rail network weight, is probably a happy medium.
There is currently no zoom visibility distinction for the various ‘types’ of railway and I was thinking that industrial and branch lines should not appear until higher zoom levels. OSM uses 13 and it seems about right while Americana is 14 and it is way too harsh (but seems to have a different coverage to OSM iD). Note: I’m no expert here and may be confusing ‘usage’ and ‘service’ tags as it could be that OSM and Americana use ‘service’ to define zoom visibility. As ‘service’ should only be applied to ‘minor lines’ (Key:service) and ‘relatively short lengths of track’ (Tag:service=spur), it makes sense that it governs visibility, but I have seen OSM examples of ‘service=spur’ that are 10 km or more on ‘usage=industrial’ lines. Zoom visibility based on ‘service’ has to be implemented (13=visible?) and probably with no distinction between its values. Perhaps zoom visibility of ‘usage’ for values other than ‘main’ needs more thought as there are others beyond industrial and branch, e.g. tourism (Key:usage). There also various values of ‘railway’ that I haven’t considered - abandoned, disused, etc. that seem more relevant in the OHM historical context (Key:railway) and they should probably have zoom visibility criteria.
@AndrewS_OHM - great to see you here again. : ) I’ll need some time to merge / incorporate those thoughts, but I very much appreciate the thorough, structured feedback!
Given their importance for European passenger transport, railways should be more visible. Currently, the main distinction is between usage=main and usage=branch (Each country has its own legal definitions for it.)
A new railway tagging scheme would be based on passenger train service.
usage=highspeed and usage=intercity (and usage=luxury) would be useful, but:
highspeed trains also operate on lowspeed or mixedspeed routes
intercity trains are sometimes discontinued for financial reasons while the track remains the same
@fk27 - This is interesting. Clearly, the trains are separate from the tracks, and track infrastructure seems to be the primary focus of much of our efforts. But, your comment has me thinking: maybe we should show routes on the Historical layer at lower zoom levels. Maybe we’d have to define what are significant intercity routes, but maybe that would obviate the questions of main vs branch?
I’m always confused when different types of things are tracked in the same key. For example: usage=highspeed is more of an attribute of speed than usage, but I’m a recent arrival to rail mapping, so there may be other implications. I could easily see where a route is marked as usage=highspeed; passenger; luxury, etc…
Exactly for the 2nd “but” you raised, modifying the railway=rail track is not scalable in OSM, as it would require creating a new object whenever it changes). route=railway depending on definition, or aggregating (detecting) route=train on them is more viable.
In any case, usage= is a conflict with other usage= , and =highspeed is confusing with how service= is used on railway= and route= ( service=high_speed ) differently. service= has quite a few problems with mixing different aspects (eg PRC high-speed overnight sleeper; Japanese regional HS short services, intercity/interregional / long-distance vs regional limited-express; too German or European centric definition, and there are already RE and IRE on top of =regional if for RB) . There’s passenger= in fact, for the geographical extent separately.
@jeffmeyer You may start with Sunset Limited (AMTRAK #1/2): a train that has changed its route several times since 1894. However, that route still needs a lot of work.
Railways have never been, under any circumstance, even in Europe, been more important than highways.
There are other tags for highspeed raillines, highspeed=yes.
are you going to edit usage=intercity with every schedule change? I think we should be talking about the tracks, you can create relations for services. What even is usage=luxury?
I would strongly advocate to keep the OSM tagging scheme for tracks intact.
, I thought I’d beg this time. 
… add a comment or two here… 0 replies is so demoralizing!