OpenStreetMap logo OpenStreetMap

Changeset When Comment
153819213

Hi, thanks for adding the opening hours of the Drawing Center.

I was changing the hours into the standard format that we in OSM (documented at opening%20hours=* but the short version is that we use 24 hour time and two-letter day abbreviations) and I noticed a discrepancy between the hours you added and those listed on the Drawing Center's website.The website says they're open Thursdays but you wrote "Thursday Closed."

I used the info from the website, but is the website wrong?

146562800

I clearly remember seeing a "no motorcycles" rule in place at one time, but I haven't been able to find it since. Was this signed at the cemetery?

141401927

Thanks for adding this path: way/1208651473

I just used it and saved a few minutes :)

152588296

(I also reopened the zoo!)

152603444

(changeset comment should have said "add address")

149875715

Howdy! On some intersection nodes eg node/11455767273 you've added crossing=traffic_signals but I'm not sure that's correct. This tag shouldn't be used unless there's a walk signal for pedestrians. As far as I know this intersection just has stop signs, not signals.

149420271

Hi isakavya -- the tag "access:conditional=no @(Mo-Fr 07:00-13:00) 07:00-13:00)" on the segment way/1268090833 that you created is badly formatted (mismatched parens).

Your changset comment "added restriction for board is only for buses" doesn't exactly make sense either -- are you trying to say that this entire block of Lexington is bus only for those hours? Perhaps you noticed a restriction sign for the bus lane and thought it applied to the entire roadway?

149223746

Thanks!

149395903

Nah they didn't have one posted at all, good catch, fixed

148569753

Great, let me know if you need any help, or eyes on the ground. Happy mapping!

148569753

Hi kguenther24!

I see you've been adding a lot of highway=pedestrian + area=yes polygons (& multipolygons) in downtown Brooklyn. Unfortunately the mapping methods you're using are causing errors in pedestrian routing. Examples:

- From Columbia Heights down to Squibb Park (osm.org/directions?engine=fossgis_osrm_foot&route=40.70077%2C-73.99523%3B40.70088%2C-73.99548#map=19/40.70089/-73.99528):
The route now follows the edge of way/1257037956 instead of taking the long ramp down into the park.

- From Sands Street to the Manhattan Bridge walkway (osm.org/directions?engine=fossgis_osrm_foot&route=40.69988%2C-73.98712%3B40.70005%2C-73.98682): The route now follows way/1262001717 (added in this changeset) and jumps from ground level directly up to the bridge. This bypasses the stairs, which is especially bad because it impacts wheelchair and bicycle routing. (Bicycles aren't actually allowed here at the moment, of course.)

Current screenshots are at https://jmapb.github.io/kguenther24/ in case the map changes before you read this.

The underlying issue seems to be that you're using highway=pedestrian + area=yes tags to describe the shape of sidewalks and other linear footways. This is incorrect tagging because area=yes should only be used on a highway area if it can be navigated omnidirectionally -- that is, when a traveler can meaningfully move in any direction and enter and exit the area at any point where it connects to other highways.

For linear (one way or two way) highways, you can use area:highway=* tags to map the area details. See area:highway=footway for sidewalks and other footways, area:highway=steps for stairs, and area:highway=pedestrian for linear pedestrian highways. These tags will allow the routing engines to process the highways as linear, which should keep these errors from happening.

Thanks, jmapb

149223746

Hi there! Looks like you disconnected the cycleway segment way/1062089087 from the crosswalks at node/6815003765 and node/11766086963. Was this intentional?

Thanks, J

147629506

Hi George, I came to finally scout this 4-year-old note note/2222572 but you beat me to the punch, congrats!

I noticed you left this small segment of Milla Drive: way/20237460 -- was that intentional? I didn't see anything that would indicate that there's any road name around here other than Sun Mountain Road.

143557288

Hi snake21, what's the source for this only_straight_on relation? relation/16640114

147794169

Well that's a fine question. I guess my instinct was to put the city department first, a la nycdoitt:bin. I've also used nycdoe:ref for schools. But I'm not married to these forms & happy to standardize on whatever folks think is best.

111873382

Is this really a no-smoking law office? ;)

node/9132534664/history/1

146577394

From a road in Ulster County, no less! I'm tempted to say it's a Go Map bug but more likely it was human error, somehow? Thanks for the heads up.

146461191

Absolutely, if possible. I only map this way as a last resort.

Here's one I remember doing a while back: way/249606377

It's signed 798-812, right under the Colonel's nose. It's 798-812 on the website ( even right in the url: https://locations.kfc.com/ny/brooklyn/798-812-4th-avenue ). There's no reason to think the restrooms are 798 4th Ave and the drive-thru window is 812 4th Ave. By all available evidence, the real-world address of this place is 798-812. So that's what I tagged, and added the address interpolation way to boot. Frankly it's hideous, but I still think it's less bad than any currently-available alternatives.

146461191

"-" really is a damned boondoggle. I really wish Queens hadn't adopted it as separator punctuation in their addresses, and I really wish other New Yorkers wouldn't use it to sign ranges. But they did, and they do.

I've struggled a lot with how to deal with this. When a poi really does appear to be using the whole range as their housenumber (2752-2758 in this example) my instinct is to angrily tag what I see.

The other obvious tactic is just to tag it with the lowest housenumber in the range. This looks better from a data hygiene point of view, and avoids any Queens-style confusion. But it conveys less information, and it doesn't follow OSM's "truth-on-the-ground" best practice.

Nonetheless, putting the whole range in addr:housenumber is annoying enough that I'd much rather just use a single number, if I can justify it somehow. The good news is, I think I can justify it here -- seems this place has a website https://www.atlanticcarsale.com and the housenumber listed there is simply 2752. So, making believe that the website is equivalent to on-the-ground signage, I've changed this car dealer to `housenumber=2752`.

I could talk about this problem all day, and probably should at some point.

If you haven't read it yet, here's NYC's summary of the address formats in use & how the city's own geocoding software handles them: https://nycplanning.github.io/Geosupport-UPG/chapters/chapterV/section02/#house-number-format-standards

And here's a proposal for a tag that might be of use to us, if it were ever adopted by geocoders: osm.wiki/Proposal:Addr:interval

142276477

Aha, thanks. I've removed all the Giovanni tagging from node/2550089888 and left it as a disused restaurant.

The website says "diagonally across the street" so I'm guessing it means at way/249624759 which has been vacant as long as I can remember. I left a note there so we can keep an eye on it.