On Feb 4, 2018, at 7:47 AM, Mike Boos <mike.b...@gmail.com> wrote:
> I've noticed some users have begun tagging some roads in a number of Canadian 
> cities with lcn=yes tags, which are intended for marking local cycling 
> routes. My understanding of the lcn tag was that it was intended for marking 
> designated routes, not just any old way that is potentially bikeable or 
> personal preferences cycling routes.

Hi Mike:

Thanks for the heads-up about your intentions.  Yes, I agree with your 
characterization that lcn=yes (either directly on a way or on a route relation 
with ways collected together as elements) is for "designated" routes.  And that 
is one way to do it:  another is if you know the "level" of either federal, 
provincial or local) where you might tag network=ncn, rcn or lcn in a bicycle 
route relation.  These routes might be designated by a government (federal, 
provincial or local) or a "cycling group" (could be a non-profit or for-profit 
member organization).  And/or they might be a proposal before a government body 
(for example, if a set of bicycle infrastructure so tagged has a numbering 
protocol which organizes that set), in which case you'd add state=proposed, 
then OpenCycleMap would render the route lines as dashed instead of solid.  But 
yes:  OSM has reached a consensus that such routes designated in our map data 
are not "simply personal preferences" by a single person who wishes to organize 
bicycle routes, whether on infrastructure tagged with bicycle-ish tags like 
highway=cycleway or cycleway=lane, or not.  They are either signed on the 
ground or perhaps published as a paper or online map by a government or bike 
non-profit who has organized infrastructure together as routes and publishes 

> For many roads, the lcn tag seems redundant, since these ways are already 
> tagged with cycleway=lane or something similar, and there is no accompanying 
> lcn_ref tag to provide information on individual route names or numbers (if 
> they exist). Other roads have been tagged, but have no infrastructure or 
> signage, which suggests someone is simply marking their personal routes.

There really are two tagging schemes going on here, for different purposes:  
one is tagging infrastructure (nothing to do with lcn or lcn_ref), like 
highway=cycleway or cycleway=lane or bicycle=yes (there are others).  The other 
is specifically for routes, and these include lcn (or much less commonly 
lcn_ref) and/or a relation with tags type=route, route=bicycle and usually a 
network=*cn (lcn, rcn, ncn, icn).  A newer (though established in both Canada 
and the USA) tag of "cycle_network" is also being applied, this helps to 
disambiguate specific bicycle networks which share the same network "level."

The lcn tag isn't redundant:  a way can be tagged as bicycle infrastructure 
(and that is one thing) independent of being part of a bicycle route.  The lcn 
tag is used when (USUALLY, perhaps ALWAYS or NEARLY ALWAYS) bicycle 
infrastructure IS part of a (local) route.  Bicycle infrastructure can be so 
tagged while NOT being part of a route, but the converse usually isn't true.

> I'd like to think I have some sort of expertise in what constitutes an 
> official local cycling route in my area, having served as a member and later 
> chair of the Kitchener Cycling and Trails Advisory Committee for several 
> years. There are some signed routes that myself and others in the area have 
> properly marked with relations. But is my understanding of what the lcn tag 
> is for wrong? I'd like to know before I start cleaning things up.

I would strongly agree with you that signed-on-the-ground routes (almost always 
part of a "network") should be mapped in OSM and that "casual" or "personal" 
routes should not be mapped in OSM.  Again, an exception can be a formal 
proposal for a network or numbering protocol which has been introduced before a 
government body and is either in the process of being approved or implemented.  
For example, routes in the USBRS, USA's national bicycle routing network / ncn 
have been emerging over the last several years and will take many more years to 
complete as the network gets "built out".

In the interests of gaining a solid understanding of this (and to "keep your 
sanity!" while learning), I recommend that you FIRST "get infrastructure 
tagging correct" on the ways which are actually bicycle infrastructure:  
highway=cycleway, cycleway=lane, and so on.  SECOND, assemble these bicycle 
infrastructure elements into route relations (type=route, route=bicycle, 
network=lcn/rcn/ncn/icn).  I speak from deep experience:  sticking to that 
ordering, while not required, really helps.  Take your time (over weeks) and 
watch Cycle Map layer render infrastructure tags a certain way and route 
relations in a certain way.  I'm not saying "tag for the renderer," I am saying 
"tag correctly (by what our wiki says) and you'll get great results rendered."

Please take a look at these wiki pages of ours:

https://wiki.osm.org/wiki/Cycle_routes (the definitive guide to routing, but 
not infrastructure)
https://wiki.osm.org/wiki/United_States/Bicycle_Networks (how the USA assigns 
routes to the various levels of national, regional, local)
https://wiki.osm.org/wiki/WikiProject_U.S._Bicycle_Route_System (the specific 
example of how far along the USA is in implementing its national cycleway 
https://wiki.osm.org/wiki/Key:cycle_network (has a section both for Canada and 
https://wiki.osm.org/wiki/WikiProject_Canada#Trans_Canada_Trail (brief, but 
https://wiki.osm.org/wiki/FR:Route_verte (in French, but how one province 
implements a regional cycleway network's routes in OSM)

In the latter, especially note the table near the end after "Voici les 
attributs recommendés les segments associés à chacun des types:"  It shows how 
all member elements of the routes have specific infrastructure tags for 
bicycles.  That's right!

If Canada would like to create a 
https://wiki.osm.org/wiki/Canada/Bicycle_Networks wiki, I'd be delighted to 

Asking questions is great, I hope I have provided you with some answers.

Talk-ca mailing list

Reply via email to