I know the ABS SUA includes the Blue Mountians, but I don't consider the Blue Mountains part of Sydney, so Emu Plains would be in Sydney but Lapstone not.

ABS Boundaries aren't administrative boundaries, rather just ABS's view for statistical reporting.

It's hard since you can't survey this on the ground, I'm not aware of any official source, so the best approach is really to consider what do most people consider part of Sydney.

What do you think?

1. Not in this jurisdiction. http://www.rms.nsw.gov.au/roads/safety-rules/road-rules/road-markings.html "Drivers are permitted to cross a single dividing line enter or leave a road".

2. I'm not talking about measuring distance for a route, this is irrespective of navigation. It's purely the fact that we no longer have a centre line representation of this non-physically separated road. Which is a loss, but you're right that practically there is not much impact.

Sorry if I sounded negative, I really do appreciate you taking the time to work through this.

Two issues are:
1. You can turn across a solid white line (which is what's separating these two lanes to the south west of the intersection) to enter or leave a property. Mapped as a single way a router can direct you to a destination on either side of the street via either direction of travel on the road. Mapped out as dual way the router must direct you to your destination via only one of those ways (since normally dual carriage ways mean there is a physical separation which you can't cross).
I guess if physical separators (barrier=fence, traffic_island, raised curb, etc) are mapped out in theory routers could try to use that to distinguish dual ways with a physical separation and those with none, but I can see that getting quite messy and to get right.
2. It makes it much harder if not impossible to measure total road length from the OSM database as this single road is no longer the road center-line but instead the center-lane-line. This is less of a concern as it's a problem with the accepted way to map physically separate ways as dual ways, so not really specific to this case.

PS. I fixed up some of the turn:lane tags in changeset/51802074

I wouldn't exactly call it solved, as turning a road which has no physical separation between the directions into a dual way road creates issues which weren't there when mapped with a single way. But I guess now we have the stop mapped. +1

In OSM we almost need a way on the road centre line representing the physical road which is tagged as a painted solid line, painted dotted line, raised barrier (which can be driven over in an emergency), plastic poles or fence, along with other details about the road line name, classification.

And then separate set of ways representing the road network for routing with ways for different directions and lanes like it's been mapped in this changeset.

Thanks for replying. I looked at it both just recently and about a year ago. Things change and imagery get's outdated. I know how easy it is to think it represents reality but it's not always the case.

Could you please explain why this was changed from bicycle=designated to bicycle=yes?

Could you please let me know what you mean my "Sourced from RMS"? I haven't heard back from you about this yet.

Hmm I always read access=* as designated meaning it's set aside for that purpose. So a path with has a sign saying pedestrians and bicycles can use it would be foot=designated, bicycle=designated. access=yes I understood as you're allowed to use the way, but there is no signage to explicitly allow it, so a typical footpath would be foot=yes since people on foot a legally allowed to walk on it, but there is no sign setting it aside specifically for pedestrians.

So I think that matches most roads being psv=yes and bus lanes as psv=designated.

In the case of cycle paths/lanes, signs indicating they are for bicycles means bicycle=designated.

Sorry I'm not sure what part you feel should be added to the wiki?

You're saying that this isn't a legal (in NSW) cycle lane because it lacks the word "LANE" painted in white next to the white bicycle symbol painted in the bicycle lane? Maybe you want something like bicycle:lane:legal=yes/no? I don't think it needs to meet the NSW legal definition as a cycle lane to be mapped as a cycle lane in OSM when it matches the OSM wiki guidance of what is a cycle lane.

Sorry I find cycleway:lane:bicycle confusing. The cycle infrastructure in question here is designated for bicycle since it has a bicycle symbol painted in the lane.

If you'd like to be able to distinguish narrow cycle lanes from wide ones please use the cycleway:lane:width tag as described at cycleway=*#Supplementary_details

There is a solid while painted line separating the traffic from the bike lane and a dotted white painted line separating the bike lane from the curb side parallel parking. These bike lanes have a painted white bike inside the lane which designates them as a bicycle lane. The bike lane can't be a shared_lane as it's not wide enough for vehicles.

I think we should go by the OSM wiki as it provides a globally consistent way of tagging rather than using local/jurisdiction specific legal definition. The wiki says "A bicycle lane...is an inherent part of the road itself" Check. "It has no physical separation from the other lanes except the painting on the road." Check. "Notably, there is no curb between the cycle lane and the road" Check.

The painted bicycle icon in the bike lane is enough, it doesn't need the word lane painted there too to be a bike lane.

I've changed some of these you tagged as shared_lone back to a cycle lane in changeset/51705715. It's marked out as a cycle lane not a shared lane on the road so should be tagged as such.

I've added the cycleway:lane=doorzone tag which might help you distinguish doorzone cycle lanes. Also some of the places you added shoulder there is no shoulder so I fixed that up.

Some were already changed back to cycle lane in changeset/33213526

I've changed cycleway=lane to cycleway=shoulder in some places which I surveyed in changeset/51705625.

Given there are no markings on the road (signs, paint on the road) indicating a cycle lane, I think the shoulder tag is more appropriate.

https://taginfo.openstreetmap.org/tags/cycleway=shoulder

It seems this changeset added a lot of maxspeed to nodes, which seems redundant when on the way. It makes editing later a lot harder too.

I don't understand why way/495578455 was added back in, did they temporarily reopen this northbound section, it seems very strange to me that they would.

I deleted the northbound road in https://osmcha.mapbox.com/changesets/46239272 6 months ago but it's been added back in in this change, I've just deleted it again in changeset/51667353

I did a Mapillary survey, but it will take about a week to upload.

Hi @pangenib. Welcome to OSM and thanks for contributing.

Could you please clarify what you mean by "Sourced from RMS"?

You might want to check out osm.wiki/Copyright#Proprietary_data. We can't use copyrighted data without a license in OSM, so that rules out including details about bicycle infrastructure from RMS maps or data without a license.

If there is doubt on the source and copyright of the data then we may need to revert your changesets, so it would be great if you could please clarify your source?

You're more than welcome to add this data from other sources like Aerial Imagery we have permission for, Mapillary, or ground surveying.

even if it translated, it's common in OSM not to add translations but only name:<lg> if the feature is known by that foreign language name. Although I'm not strongly one way or the other.

Hi, Welcome to OSM and thanks for contributing.

In the case of way/399938817/history the addr:street tag is reserved for the street address. This roof doesn't have an address so it shouldn't have an addr:street tag.

Hi, Thanks for contributing to OSM!.

You can read up on names at osm.wiki/w/index.php?title=Key:name&uselang=en-AU.

name= should be the name the sign says, if it also has a name in another language you can add this with something like name:zh

I think we need to have a bit more of a discussion on this. If people want the LPI Admin Boundaries they can go to LPI and get it. The LPI boundaries aren't the official legal definition of these boundaries, they are just the LPIs representation.

I think we should do what makes sense for OSM and not necessarily just copy LPI.

I thought quite often admin boundaries are defined in reference to a road, coastline or river. In those cases I do think that the admin boundary relation should use the road, coastline or river way as part of it, if that's it's definition.