cycleway=lane only needs to have a marked (painted) separation from vehicles, whether that bicycle lane is in a door zone or not is irrelevant to if it's a cycle lane or not per the OSM definition. ie. being in a door zone does not all of a sudden mean it's not a bicycle lane.

The tagging list discussion on how best to tag a door zone is at https://lists.openstreetmap.org/pipermail/tagging/2020-May/052328.html and my preference from that discussion is cycleway:lane:doorzone=yes. I'm working to put this through a formal proposal process so it can be an approved tag, then data consumers can look for this tag. In the meantime this should still be counted as a bicycle lane.

Since cycleway=doorzone is has not really been discussed or gone through a proposal process, I feel we should still use cycleway=lane so data consumers can still make use of this. cycleway:lane=doorzone has been used in Australia to mark cyclelanes which are in a doorzone but now this conflicts with cycleway:lane=* so we need a new tag, but to replace cycleway=lane with cycleway=doorzone I think needs to be discussed.

I've reverted your change for now.

When I surveyed this a month ago there was still some evidence that the used to be a track here, so I marked it using the lifecycle prefix osm.wiki/Lifecycle_prefix as an abandoned track "Still visible but fallen into serious disrepair and which could only be put back into operation with considerable effort". Unless it's completely regrown with nothing left visible on the ground, it wouldn't hurt to leave it here with a suitable lifecycle prefix.

What happened to the elevator?

I've removed it now then.

As mentioned on talk-au, lcn=yes is valid on ways which make up a route. The network tag can go on a relation, but I don't think it should go on the way.

node/7443163061/history is there anything on the ground here?

I've made a few changes, see https://www.mapillary.com/app/?focus=photo&pKey=-cAmRjhSGlVIYsIbZxyAlg&lat=-33.750527777777776&lng=151.23233055555556&z=17&x=0.9297334205115395&y=0.5080898438599728&zoom=0 the building doesn't extend out like you've mapped it, it's just a roof and not part of the building, so I've removed that. Same of the covered footway on the north. Also I've marked the inside part of the building as building:part=yes.

I've also reverted this changset since it removed buildings which were recently constructed and likely the imagery traced here is outdated. See Mapillary imagery for more recent survey.

I've reverted this changeset, area has been undergoing changes and construction and sources may not reflect what's currently on the ground.

I've reverted this changeset, area has been undergoing changes and construction and sources may not reflect what's currently on the ground.

I took a look at the aerial imagery so can see what this one is now, it's like the Kent Street cycleway https://www.mapillary.com/map/im/lkVf4biZTxTfMwb57PxUcg.

These appear to match "cycle tracks" listed at osm.wiki/Bicycle#Cycle_tracks

That section of the wiki says there are two alternatives to map, a single way or separated ways. Both are valid, but where someone has taken the time and effort to map it out separately that shouldn't be removed in my opinion.

osm.wiki/Bicycle#Cycle_features also describes the two approaches and notes the different opinions on which approach to use.

While mapping it out as a separate way can make it harder to link it back to the road segment and hence as you point out some routing engines don't do as good a job, it does make it easier to tag things like the width, surface, traffic calming, turn restrictions as a separate way and more accurate geometry.

Probably best to also take this discussion to talk-au so we can determine a consensus and best approach.

> It's more or less the same reason as to why we don't go around mapping every footpath adjacent to a road as a separate way.

Actually I do map it as a separate way and many others do to this too, it makes for much for accurate representation of what's on the ground, ease of tagging attributes and better routing especially around crossings.

Hi I think this probably should be highway=pedestrian + area=yes ?

If the cycleway is physically separated from the road then it's better to map it as a separate way as highway=cycleway. If you're short on time it can be mapped as cycleway=track but if someone else has mapped as a separate way it should be left. Unless there is no physical separation from the road in which case it should be cycleway=lane. Which is this?

looks like you've accidentally dragged the address node here

Neither is wrong, just different styles, in fact I would say how you had it before is better, but still either way is acceptable.

Sounds good thanks for confirming. There's probably different opinions on what's trail_visibility=no and what's trail_visibility=horrible, for me it's generally no being completely off-track and horrible being some evidence of a path but frequently disappears. I'll trust your best judgement, it's more important to have no/horrible than none at all and someone things it's a well marked out track.

Yeah I saw your edits adding route=canyoning, good to see someone interesting in improving OSM in this aspect.

See also natural=gorge you can use to mark the natural canyon feature (if it's a wet canyon then the tag would be on the same way as waterway=stream). I've done it for a dry canyon at way/732287179

Eventually I start including natural=gorge tags on my maps for beyondtracks.com so would be nice to see more of these tags added.

route=canyoning normally would go on the relation not the way at least following osm.wiki/Proposed_features/Canyoning but I'm not going to complain about it going on the way itself, it's not harmful and can always be converted to a relation later.

I hope you're not copying Tom's maps but instead using your own surveyed data since his maps are copyright, that said he does edit here sometimes, but you'd need to check and make sure he's happy to license it for use by OpenStreetMap.

Hi I see you've added trail_visibility=no, that usually only applies when there is no path at all is that the case here?

reading some forums seems like plenty of people just ignored the sign and ride here anyway, but also looks like it's within the last few years it changed. Either way I've updated it now, but we should try and check some of the other ones you've added here.