Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
The tag oneway=true is extinct in the database. Without defending the author of the Craigslist stylesheet: tracking OSM data changes is hard. In part due to the negative attitude towards cleanup mechanical edits, the data is all over the place. 1, -1, reverse, true, false, no, yes, maybe. Aghgh. Stuff gets deprecated then rots in place for weeks months years. How could a rendering engine keep up with all the positive and negative terms people dream up, wikifi, and then use inconsistently? - Mechanical edits help data consumers. Clear tag descriptions help data consumers and mappers. Editor validations helps data consumers, and it's rather weak right now. API level rejection of certain edit patterns would help even more. A clear migration path between tagging schemes would help data consumers a lot. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
just a few thoughts: What is the value of a 1 time mechanical edit cleanup ? From the moment you ran your script, new data can arrive in the OSM with the wrong values. Will you run your script daily ? What if a data consumer obtains the data between 2 runs of your script ? Wouldn't it be better to properly inform the data consumer about the existence of some fix scripts that can be run after obtaining the data from OSM ? Of course one could imaging an OSM setup where those mechanical edits would run automatically after each submission, but that is out of the question I think. Part of that work could be (or is) done by the validators in the editors. regards m On Mon, Jun 1, 2015 at 8:36 AM, Bryce Nesbitt bry...@obviously.com wrote: The tag oneway=true is extinct in the database. Without defending the author of the Craigslist stylesheet: tracking OSM data changes is hard. In part due to the negative attitude towards cleanup mechanical edits, the data is all over the place. 1, -1, reverse, true, false, no, yes, maybe. Aghgh. Stuff gets deprecated then rots in place for weeks months years. How could a rendering engine keep up with all the positive and negative terms people dream up, wikifi, and then use inconsistently? - Mechanical edits help data consumers. Clear tag descriptions help data consumers and mappers. Editor validations helps data consumers, and it's rather weak right now. API level rejection of certain edit patterns would help even more. A clear migration path between tagging schemes would help data consumers a lot. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
2015-06-01 9:55 GMT+02:00 Marc Gemis marc.ge...@gmail.com: Part of that work could be (or is) done by the validators in the editors. +1 If you look at the actual values for oneway, very few are obviously mistagged, e.g. oneway=yes;no (32) and no;yes (111) - these result very likely from merging streets with different properties. I have tested this with current JOSM, if you combine a way with oneway=yes and one with oneway=no you will get to a window which asks to choose from yes, no, /none/, /all/ and you have to deliberately choose all to get an ambiguous value. iD is working similarly, but it doesn't let you decide, it automatically creates a double value (yes;no), and the direction is taken from the first way selected (e.g. 1st way selected is oneway=no and 2nd way is oneway=yes and pointing in the opposite direction will revert the direction of the oneway=yes way without notice and will create a oneway value of no;yes also without notice. I have opened a ticket for this behaviour: https://github.com/openstreetmap/iD/issues/2670 ) cheers, Martin ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
On Mon, Jun 1, 2015 at 12:55 AM, Marc Gemis marc.ge...@gmail.com wrote: just a few thoughts: What is the value of a 1 time mechanical edit cleanup ? From the moment you ran your script, new data can arrive in the OSM with the wrong values. Will you run your script daily ? What if a data consumer obtains the data between 2 runs of your script ? In my experience, two rounds of mechanical edit are sufficient. The first gets the bulk of the problem, and generally shifts the mapping behavior as well. Then maybe a few weeks later check if anyone is using the old tags and reach out to them via private message. The value of a 1 time edit is high, as long as there is sufficient consensus or apathy about the change. *The potential for data consumer impact seems highly overrated on this mailing list. *By the time something is mechanically edited data consumers have generally moved on. See also http://wiki.openstreetmap.org/wiki/Deprecated_features For example take amenity=dog_bin. That's a minor tag, and chances are any data consumer also processes amenity=dog_waste_bin. Thus retagging dog_bin - dog_waste_bin would have *zero data consumer impact no matter when the script is run.* --- The tag migration is not the issue really. The problem comes in when the old tag was semantically unclear and can't be migrated. I ran into that with excrement tags actually: one set of excrement tags was used for *four distinct features* and to clean that up I had to look carefully at the context and/or ask a local mapper to verify on the ground. Fortunately the context made the right tag clear in most of the cases: but it could have been really hard. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
Yeah, I noticed this about the Craigslist rendering a while ago. In my case it was ways tagged as oneway=-1. Since the ways were not members of any relations and there was no other reason for them to be tagged this way, I reversed the way directions and changed the tag to oneway=yes but obviously this doesn't apply to the situation that started this thread. FYI: last time I checked, Craigslist also had random problems rendering water. Some lakes just don't show up in their map even though they are tagged correctly and don't have any tricky geometry. This has generated more than a few notes. I made a comment about this in the map feedback forum and one of their devs hopped on to IRC the next day and said that they were aware of it but hadn't figured out what the problem was. Toby On Mon, Jun 1, 2015 at 4:35 AM, Martin Koppenhoefer dieterdre...@gmail.com wrote: 2015-06-01 9:55 GMT+02:00 Marc Gemis marc.ge...@gmail.com: Part of that work could be (or is) done by the validators in the editors. +1 If you look at the actual values for oneway, very few are obviously mistagged, e.g. oneway=yes;no (32) and no;yes (111) - these result very likely from merging streets with different properties. I have tested this with current JOSM, if you combine a way with oneway=yes and one with oneway=no you will get to a window which asks to choose from yes, no, /none/, /all/ and you have to deliberately choose all to get an ambiguous value. iD is working similarly, but it doesn't let you decide, it automatically creates a double value (yes;no), and the direction is taken from the first way selected (e.g. 1st way selected is oneway=no and 2nd way is oneway=yes and pointing in the opposite direction will revert the direction of the oneway=yes way without notice and will create a oneway value of no;yes also without notice. I have opened a ticket for this behaviour: https://github.com/openstreetmap/iD/issues/2670 ) cheers, Martin ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
Yes, I would class this as a rendering error on Craigslist's part, not a data error on OSM's part. It sounds like the code is checking for the presence of the oneway tag, where it should be checking for the presence of oneway=yes. -- John F. Eldredge -- j...@jfeldredge.com Darkness cannot drive out darkness; only light can do that. Hate cannot drive out hate; only love can do that. -- Martin Luther King, Jr. On May 31, 2015 2:05:06 AM Bryce Nesbitt bry...@obviously.com wrote: After a long stretch of oneway=yes, I might indeed tag oneway=no just to keep someone from assuming I'd made a mistake. oneway=no is a declaration, as opposed to a lack of information. -- ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
On 1/06/2015 6:42 AM, John Eldredge wrote: Yes, I would class this as a rendering error on Craigslist's part, not a data error on OSM's part. It sounds like the code is checking for the presence of the oneway tag, where it should be checking for the presence of oneway=yes. Or oneway=1, -1 reverse, true, reversible On May 31, 2015 2:05:06 AM Bryce Nesbitt bry...@obviously.com wrote: After a long stretch of oneway=yes, I might indeed tag oneway=no just to keep someone from assuming I'd made a mistake. oneway=no is a declaration, as opposed to a lack of information. Is not the note= key to be used for passing that kind of information to other mappers? To me the note= key is much more indicative to fellow mappers than a value in another key. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
Warin 61sundowner at gmail.com writes: Or oneway=1, -1 reverse, true, reversible The tag oneway=true is extinct in the database. -- Andrew ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
After a long stretch of oneway=yes, I might indeed tag oneway=no just to keep someone from assuming I'd made a mistake. oneway=no is a declaration, as opposed to a lack of information. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
On Sun, 2015-05-31 at 07:56 +0200, Marc Gemis wrote: I've seen this topic being discussed here or elsewhere in the past. I thought that the consensus was that in some area's (a Spanish town I believe), it was ok to leave the oneway=no. The reasoning was that most streets in that town were oneway=yes, and the oneway=no was used to indicate that this road was surveyed and and an exception to the oneway=yes. so a world-wide mechanical edit should not be performed. +1 iD does prompt users to fill in tags such as oneway=no, horse=yes on roads and does I think lead new mappers to believe that such tags are required. Whilst I sometimes remove these as part of an edit, a mechanical edit would be wrong for the reasons stated above. Phil (trigpoint) ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
Forgot reply to list :( On 31/05/15 01:51, Mike Thompson wrote: Why care? I would imagine to the majority of the users make no distinction between OpenStreetMap and Craigslist's rendering of OpenStreetMap, if the map is wrong, to them that simply means that OpenStreetMap is wrong. There are a number of reasons why oneway=no is a valid tag, so if the renderer is simply reading 'oneway' and ignoring the =no one has to ask the question how many other similar tags are also being incorrectly handled. In this case Craigslist is simply wrong and there is nothing openstreetmap can do about that ... or would want to do about it on the data! -- Lester Caine - G8HFL - Contact - http://lsces.co.uk/wiki/?page=contact L.S.Caine Electronic Services - http://lsces.co.uk EnquirySolve - http://enquirysolve.com/ Model Engineers Digital Workshop - http://medw.co.uk Rainbow Digital Media - http://rainbowdigitalmedia.co.uk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
Mike Thompson miketho16 at gmail.com writes: The issue is that some streets that are not one way in OSM data, are rendered as one way by Craigslist. If you go to this ad and zoom in on the map you will see the arrows designating one way streets, for example on Cliffrose Way (https://www.openstreetmap.org/way/170048276): https://fortcollins.craigslist.org/gms/4991940913.html Here is the approximate area on the default OSM map: https://www.openstreetmap.org/#map=17/40.51930/-104.85951 I looked at the data and all of the streets in question seem to be tagged oneway=no. Admittedly this tag is rather superfluous in these cases - or at least I believe it is, I am not the one that added it, but it shouldn't be interpreted as indicating a one way street. Does Craigslist also get oneway=-1 (traffic opposite to the direction of the way) wrong? -- Andrew ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
On 31/05/15 12:24, Andrew Hain wrote: Does Craigslist also get oneway=-1 (traffic opposite to the direction of the way) wrong? And oneway:cycle=no and the like ... -- Lester Caine - G8HFL - Contact - http://lsces.co.uk/wiki/?page=contact L.S.Caine Electronic Services - http://lsces.co.uk EnquirySolve - http://enquirysolve.com/ Model Engineers Digital Workshop - http://medw.co.uk Rainbow Digital Media - http://rainbowdigitalmedia.co.uk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
On 31/05/2015 10:51 AM, Mike Thompson wrote: I have posted this issue to the Craigslist feedback forum, but I also thought I would post it here to see if anyone has a specific contact at Cragislist that handles their map rendering. They use - and credit - OpenStreetMap (which is wonderful), but I believe they perform their own custom rendering, which seems to be the source of this issue. The issue is that some streets that are not one way in OSM data, are rendered as one way by Craigslist. If you go to this ad and zoom in on the map you will see the arrows designating one way streets, for example on Cliffrose Way (https://www.openstreetmap.org/way/170048276): https://fortcollins.craigslist.org/gms/4991940913.html Here is the approximate area on the default OSM map: https://www.openstreetmap.org/#map=17/40.51930/-104.85951 I looked at the data and all of the streets in question seem to be tagged oneway=no. Admittedly this tag is rather superfluous in these cases - or at least I believe it is, I am not the one that added it, but it shouldn't be interpreted as indicating a one way street. I could remove this tag, but I feel this would be tagging for the renderer (actually someone else's renderer), and there may be legitimate cases for using oneway=no on a way representing a street somewhere else. Why care? I would imagine to the majority of the users make no distinction between OpenStreetMap and Craigslist's rendering of OpenStreetMap, if the map is wrong, to them that simply means that OpenStreetMap is wrong. I'd remove it. Redundant information - data base bloat. Like vehicle=yes on a motorway ... Perhaps these could be candidates for a mechanical clean up? ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Craigslist OpenStreetMap Rendering Issue
I've seen this topic being discussed here or elsewhere in the past. I thought that the consensus was that in some area's (a Spanish town I believe), it was ok to leave the oneway=no. The reasoning was that most streets in that town were oneway=yes, and the oneway=no was used to indicate that this road was surveyed and and an exception to the oneway=yes. so a world-wide mechanical edit should not be performed. regards m On Sun, May 31, 2015 at 3:22 AM, Warin 61sundow...@gmail.com wrote: On 31/05/2015 10:51 AM, Mike Thompson wrote: I have posted this issue to the Craigslist feedback forum, but I also thought I would post it here to see if anyone has a specific contact at Cragislist that handles their map rendering. They use - and credit - OpenStreetMap (which is wonderful), but I believe they perform their own custom rendering, which seems to be the source of this issue. The issue is that some streets that are not one way in OSM data, are rendered as one way by Craigslist. If you go to this ad and zoom in on the map you will see the arrows designating one way streets, for example on Cliffrose Way ( https://www.openstreetmap.org/way/170048276): https://fortcollins.craigslist.org/gms/4991940913.html Here is the approximate area on the default OSM map: https://www.openstreetmap.org/#map=17/40.51930/-104.85951 I looked at the data and all of the streets in question seem to be tagged oneway=no. Admittedly this tag is rather superfluous in these cases - or at least I believe it is, I am not the one that added it, but it shouldn't be interpreted as indicating a one way street. I could remove this tag, but I feel this would be tagging for the renderer (actually someone else's renderer), and there may be legitimate cases for using oneway=no on a way representing a street somewhere else. Why care? I would imagine to the majority of the users make no distinction between OpenStreetMap and Craigslist's rendering of OpenStreetMap, if the map is wrong, to them that simply means that OpenStreetMap is wrong. I'd remove it. Redundant information - data base bloat. Like vehicle=yes on a motorway ... Perhaps these could be candidates for a mechanical clean up? ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk