|
No, don't tag these as lcn. LCN is for a coherent network around a particular locality. Semantically "route=bicycle" in a relation is enough to say that this is a cycle route. (There may be a case for an additional tag to indicate recreational purpose.) Unfortunately lots of people do tag isolated routes as LCN simply to get them to render on OpenCycleMap (in fact I think I did that once, years ago). But they - we - shouldn't. We should encourage wider renderer support for waymarked routes that are not part of a country's National Cycle Network, an area's Regional Cycle Network, or a locality's Local Cycle Network. In the UK this would famously encompass the National Byway leisure route which is extensively tagged on OSM but not yet visible in any mainstream render. These aren't generally included in relations, just as individual roads aren't. Is highway=cycleway enough to say that it's a cycle route?
(19 Jul '10, 20:45)
NE2
In the UK, standard practice is that relations are used for named recreational trails. This might be because of the UK's general retardedness in any form of infrastructure that isn't meant for the car; such trails often tend to be 'bitty', so grouping them in a relation appeals rather than retagging endless segments. It also solves the namespace problem of two trails sharing the same path. You can also use route=bicycle as a standalone tag on a way. highway=cycleway indicates only a piece of infrastructure, not whether it's part of a continuous trail.
(20 Jul '10, 09:52)
Richard ♦
I'm not sure what you mean by this: "highway=cycleway indicates only a piece of infrastructure, not whether it's part of a continuous trail." highway=cycleway means it's continuous from one end to the other, does it not? That's what these are - long pieces of cycleway that are not part of something larger.
(20 Jul '10, 11:43)
NE2
1
We may need to break out the word "cycleway" here. One definition is "a strip of tarmac [or whatever] intended for use by bicycles". This is what "highway=cycleway" means. It simply refers to the on-the-ground infrastructure. At the other end of the scale, a 400-mile route might be described as a "cycleway". In the UK, for example, we have the "Pennine Cycleway", a recreational route combining lots of individual sections of highway, cycleway and other types of path. This would typically be tagged with a "type=route;route=bicycle" relation.
(20 Jul '10, 11:53)
Richard ♦
These are usually trails built on former railway grades. A few are marked as parts of the BicyclePA (rcn) network, but most only have individual names that apply solely to the trails, not to any connections to other trails.
(20 Jul '10, 12:30)
NE2
Ok. In that case you could either tag the ways as "highway=cycleway;route=bicycle;name=The Big Frog Trail", or put them in a relation with "type=route;route=bicycle;name=The Big Frog Trail". My personal preference would be for the latter for consistency with other recreational cycle routes, but YMMV. One advantage of using a relation is that the trail may not have a single identifiable name that can be chosen for the way: for example, last weekend I cycled on something that was both the former Somerset & Dorset Railway and the Collier's Way, and known to locals as both.
(21 Jul '10, 10:17)
Richard ♦
I'm still confused about what route=bicycle means on a way. says it can be used on ways, but the linked page only seems to talk about using it on relations. How is it different from, say, bicycle=designated (which is implied by highway=cycleway)?
(21 Jul '10, 20:34)
NE2
Very very roughly 'route=bicycle' on a way means 'this is part of a waymarked cycle route', in the same way that it does on a relation. It was principally used before relations were introduced to OSM, and has since become less popular on ways. As an example, I'd personally use 'route=bicycle' for the type of recreational trail you're talking about (though I'd use it on a relation rather than a way), together with a highway tag to indicate the physical infrastructure; but 'highway=cycleway' alone on, say, a short concrete path signposted for bikes in an urban area.
(22 Jul '10, 11:11)
Richard ♦
These aren't "waymarked cycle routes" though; they're more akin to ordinary rural highways that don't allow motorized vehicles. These are typical signs at entrances to trails: 1 2 3
(22 Jul '10, 12:59)
NE2
The name signs are sufficient waymarking that I'd consider those as suitable candidates for a route relation. I can't really tell you any more; you can make your own decision.
(22 Jul '10, 14:00)
Richard ♦
showing 5 of 10
show 5 more comments
|
|
Not all cycleways belong to a network. For example, Portland, Oregon has a lot of cycleways through parks for a block or two that aren't a part of the Portland or Metro (both local) cycleway networks. When tagging for references, the preferred method for cycleways is to use relations, as cycleway network routes and the underlying ways are usually widely varied. Support for rendering cycleway relations is already in the cyclemap renderer. The question is whether long cycleways create their own network in the absence of an otherwise-defined local network. Pennsylvania has some statewide routes (rcn), but most trails are not included in this network.
(17 Jul '10, 21:30)
NE2
|
|
Many cycleways are not continued in more urban areas because there are often residential streets capable of cycling. As a result many cycle networks may appear incomplete on a map while it is indeed quite complete. In rural areas there might be the idea that a cycleway should be built but has not yet been the priority. But this does not mean that the cycleways are not part of the cycle network. Even residential roads and alike can be signed as part of a cycle network and should have the lcn tags. The lcn tags is not exclusive to cycleways. I'm not sure that you understand the question. These are not parts of a marked network of cycle routes. They are simply named cycleways, just as one has a named road.
(17 Jul '10, 13:54)
NE2
|
To me the only questionable part in that would be the abbreviation of cycleway's name to form new cycle network with just one way. Though I'm not familiar with such trails, since it's named cycleway lcn=yes seems proper; I don't think cycle network needs to be fully connected even if fully mapped, you could have links between cycleways that are part of the network while the links themselves are not.
If lcn=yes is proper for all cycleways, then highway=cycleway should probably imply lcn=yes.
I think whether cycleway implies lcn=yes depends on location. In London it maybe makes sense as cycleways are more of an exception so every cycleway could be considered part of the network. In Finland where practically all pedestrian paths are shared cycleways it would make more sense to tag only specific routes.
tko: absolutely not. In London we have a signed local cycle network (both on an off road), that is what defines whether we use the lcn tags. There are many cycleways in London that are not part of cycle routes.