Re: [OSM-talk] A warning about gates and other barriers
El 21/09/2010, a las 20:34, Frederik Ramm escribió: Hi, sergio sevillano wrote: thats a good idea lets put access a mandatory tag (not just useful) along barrier=gate There are no mandatory tags. don´t be picky :) im sure everyone understood me mandatory as bridge or tunnel should have layer tag. we can call it reminder instead of error on validator software ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 20 September 2010 21:48, Frederik Ramm frede...@remote.org wrote: Firstly, what is behind the gate differs depending on your location. Secondly, the way behind the gate may well be reachable by other means (i.e. a detour) - it is easy to imagine a gate where vehicles cannot pass, but still vehicles are allowed on both sides of the gate! There's an example of this near where I used to live. A developer built a new residential road, which got a lot more traffic than intended because it cut the distance out to the main streets for a lot of people. So he put a gate across the middle, and made two dead end roads. You can visit either side of the gate, you just can't go through. He left it as a gate instead of blocking it off completely so that it can be opened in an emergency, but it's never been opened that I've ever known. Every map I've every seen shows it as two separate roads, and ignores the gate. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Richard Fairhurst richard at systemed.net writes: This is because a gate with no access tags implies that nothing can go through. Where on earth do you get that idea from? barrier=gate states that there's a gate. The thing about gates, as opposed to (say) walls, is that you can open them to get through. Agreed. And when mapping I tag barrier=gate if there is a gate. However, if you were writing a routing program, it would be unwise to assume that you can drive through a barrier=gate if no additional access is specified. Often a gate is locked shut. -- Ed Avis e...@waniasset.com ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 21 September 2010 18:38, Ed Avis e...@waniasset.com wrote: However, if you were writing a routing program, it would be unwise to assume that you can drive through a barrier=gate if no additional access is specified. Often a gate is locked shut. You can only make that assumption for your area, however as Liz points out, in other places the gate is for animal control, not for limiting access. In other cases the gates are opened at sunrise and shut at sunset. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
2010/9/21 John Smith deltafoxtrot...@gmail.com: On 21 September 2010 18:38, Ed Avis e...@waniasset.com wrote: However, if you were writing a routing program, it would be unwise to assume that you can drive through a barrier=gate if no additional access is specified. Often a gate is locked shut. You can only make that assumption for your area, however as Liz points out, in other places the gate is for animal control, not for limiting access. In other cases the gates are opened at sunrise and shut at sunset. Yes, these are all common situations in Europe as well, that's why we should strongly advice to add additional tags to gates. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
El 21/09/2010, a las 17:02, M∡rtin Koppenhoefer escribió: 2010/9/21 John Smith deltafoxtrot...@gmail.com: On 21 September 2010 18:38, Ed Avis e...@waniasset.com wrote: However, if you were writing a routing program, it would be unwise to assume that you can drive through a barrier=gate if no additional access is specified. Often a gate is locked shut. You can only make that assumption for your area, however as Liz points out, in other places the gate is for animal control, not for limiting access. In other cases the gates are opened at sunrise and shut at sunset. Yes, these are all common situations in Europe as well, that's why we should strongly advice to add additional tags to gates. thats a good idea lets put access a mandatory tag (not just useful) along barrier=gate so editing soft always ask for it and error software (as keep_right) highlights it as missing info. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 21.09.2010 18:17, sergio sevillano wrote: El 21/09/2010, a las 17:02, M∡rtin Koppenhoefer escribió: 2010/9/21 John Smithdeltafoxtrot...@gmail.com: On 21 September 2010 18:38, Ed Avise...@waniasset.com wrote: However, if you were writing a routing program, it would be unwise to assume that you can drive through a barrier=gate if no additional access is specified. Often a gate is locked shut. You can only make that assumption for your area, however as Liz points out, in other places the gate is for animal control, not for limiting access. In other cases the gates are opened at sunrise and shut at sunset. Yes, these are all common situations in Europe as well, that's why we should strongly advice to add additional tags to gates. thats a good idea lets put access a mandatory tag (not just useful) along barrier=gate so editing soft always ask for it and error software (as keep_right) highlights it as missing info. +0.5 I would say, it's NOT an error to not map the access tags itself - sometimes the mapper don't know this facts - as mentioned in this thread before. But I totally agree, that it's a mandatory tag to be sure at usage, so it should be displayed as error in error tracking software (keep right etc.), and mentioned as a warning (or kind of soft error) in editor software (like the josm validator). regards Peter ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Hi, sergio sevillano wrote: thats a good idea lets put access a mandatory tag (not just useful) along barrier=gate There are no mandatory tags. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 22 September 2010 02:38, Peter Wendorff wendo...@uni-paderborn.de wrote: I would say, it's NOT an error to not map the access tags itself - sometimes the mapper don't know this facts - as mentioned in this thread before. But I totally agree, that it's a mandatory tag to be sure at usage, so it should be displayed as error in error tracking software (keep right etc.), and mentioned as a warning (or kind of soft error) in editor software (like the josm validator). I've been doing what others do, tagging the access on the other side of the barrier, not the barrier itself, and a missing access tag usually means there is no access limitations. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
2010/9/20 Steve Bennett stevag...@gmail.com: Personally, I think I prefer the default that access is open to anyone who could be on both sides. But anyway, could we try and reach consensus, and then document that? beside simple access-restrictions there are also lots of cases where the gates are open at some time and closed at other times. Generally a gate is either open or closed, so that the mere presence of a gate gives the reader of the map a hint, that passage might be obstructed. I prefer asuming that a gate is closed (I tag them with access=private in these cases anyway) or you have to indicate allowed traffic (foot=yes, bicycle=yes, etc.)). Why do we have to have a default? Defaults don't work because you can't differentiate between information not present and access=default, and therefore I recommend to tag explicitly in order to get unambiguous data. cheers, Martin cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Hi, M?rtin Koppenhoefer wrote: I prefer asuming that a gate is closed (I tag them with access=private in these cases anyway) or you have to indicate allowed traffic (foot=yes, bicycle=yes, etc.)). Why do we have to have a default? The default should be applied by the software evaluating the data. The default is not an OpenStreetMap default but an application default. A conservative routing engine might assume an unspecified gate to always be closed, while others might simply apply a slight penalty for a route containing a gate, or even assume it is alway open. Bye Frederik ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 20.09.2010 10:12, Frederik Ramm wrote: Hi, M?rtin Koppenhoefer wrote: I prefer asuming that a gate is closed (I tag them with access=private in these cases anyway) or you have to indicate allowed traffic (foot=yes, bicycle=yes, etc.)). Why do we have to have a default? The default should be applied by the software evaluating the data. The default is not an OpenStreetMap default but an application default. A conservative routing engine might assume an unspecified gate to always be closed, while others might simply apply a slight penalty for a route containing a gate, or even assume it is alway open. Here we come to a great possibility for software to contribute to the OSM data. Wherever possible the software should provide a mechanism to add the data, if needed. Of course a conservative setting would avoid gates for being sure, but even that could be adjustable for the user: I have time, it doesn't matter to turn around at wrong data - but I will fix it then. I think, in future we need end user software with the ability to edit the OSM, because mapping new stuff is much easier and makes more fun than to fix small errors. talk@openstreetmap.org Peter ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 20/09/2010 09:07, M∡rtin Koppenhoefer wrote: I prefer asuming that a gate is closed (I tag them with access=private in these cases anyway) This is a bad assumption to have. Just because a gate is closed, it doesn't mean your not allowed access. barrier=gate means it's a gate nothing else. It could be open/closed all sorts of traffic could legitimately have access. or you have to indicate allowed traffic (foot=yes, bicycle=yes, etc.)). Why do we have to have a default? Defaults don't work because you can't differentiate between information not present and access=default, and therefore I recommend to tag explicitly in order to get unambiguous data. Tag what you see on the ground. If you don't know, don't tag it. Dave F. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On Mon, 20 Sep 2010 11:21:21 +0100 Dave F. dave...@madasafish.com wrote: On 20/09/2010 09:07, M∡rtin Koppenhoefer wrote: I prefer asuming that a gate is closed (I tag them with access=private in these cases anyway) This is a bad assumption to have. Just because a gate is closed, it doesn't mean your not allowed access. So if you visit lots of places in ¨outback¨ Australia, you get to open and close the gates as you go. The gate should be closed, and you are free to pass, but have to open the gate, pass the boundary and close the gate again. The point is that tags should only accommodate one piece of information. A gate is a gate, access is something else. Defaults cause confusion and arguments - we should have a system in which information is specified and not assumed. It´s cold here and I´m going to sit back and wait for the flames to warm me up. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Am 20.09.2010 12:33, Elizabeth Dodd: The point is that tags should only accommodate one piece of information. A gate is a gate, access is something else. Completely true. If a way has restricted access (and that may be physically enforced by the used of a closed gate) I tag the access restriction on the way behind the gate and not the gate. I don't see any added benefit by adding the access to a way behind a gate to the gate (!) itself. Claudius ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 20/09/2010 10:31, Peter Wendorff wrote: Here we come to a great possibility for software to contribute to the OSM data. Wherever possible the software should provide a mechanism to add the data, if needed. How? The original point of this thread was that routing software couldn't distinguish it. Dave F. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On Mon, Sep 20, 2010 at 01:21:49PM +0200, Claudius wrote: Am 20.09.2010 12:33, Elizabeth Dodd: The point is that tags should only accommodate one piece of information. A gate is a gate, access is something else. Completely true. If a way has restricted access (and that may be physically enforced by the used of a closed gate) I tag the access restriction on the way behind the gate and not the gate. I don't see any added benefit by adding the access to a way behind a gate to the gate (!) itself. This will be right for most cases, but not always. Sometimes access through the gate may be restricted, but free on both sides of the gate. An example: a service area at a highway. For regular traffic it is available only from the highway, but there are gates (normally closed) to local/service ways outside. Only authorized may open and pass the gates, though the roads on neither side is officially closed for traffic. Greets, Jacek ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Hi, Claudius wrote: The point is that tags should only accommodate one piece of information. A gate is a gate, access is something else. Completely true. If a way has restricted access (and that may be physically enforced by the used of a closed gate) I tag the access restriction on the way behind the gate and not the gate. I don't see any added benefit by adding the access to a way behind a gate to the gate (!) itself. We both agree on the a gate is a gate, but a gate does not necessarily mean that a way has restricted access. Firstly, what is behind the gate differs depending on your location. Secondly, the way behind the gate may well be reachable by other means (i.e. a detour) - it is easy to imagine a gate where vehicles cannot pass, but still vehicles are allowed on both sides of the gate! Bye Frederik ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
El 20/09/2010, a las 12:33, Elizabeth Dodd escribió: On Mon, 20 Sep 2010 11:21:21 +0100 Dave F. dave...@madasafish.com wrote: The point is that tags should only accommodate one piece of information. A gate is a gate, access is something else. Defaults cause confusion and arguments - we should have a system in which information is specified and not assumed. hi im the one who wrote that barrier implied access=no back in the day. but when the barrier tag was created i thought that that implication was forcing to always tag access along barrier. (that was a common effort and no one else until now thought that this was wrong neither me) also i never thought about routing softwares now, after this years of mapping, i think that defaults are a wrong way to tag if you have access tag, use it either way. barrier=gate alone, inmo only means there is a gate, nothing else. if there is no access tags there is no access info, period. a mapper can see a gate but has no time to stop and check if the gate is closed, locked or has opening_hours, so he should map it so the info doesn't gets lost and let other with more time tag access. same with satellite mapping, if i map a highway and i don't know what type is it, i must use highway=road and of course i don't know if there is access limits. until someone with more info than me tags it. the assumptions that routing software must take should be done by common sense of coders and explained in their own pages so the routing software user knows what to expect. It´s cold here and I´m going to sit back and wait for the flames to warm me up. me too cheers sergio ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Message: 9 Date: Mon, 20 Sep 2010 20:33:58 +1000 From: Elizabeth Dodd ed...@billiau.net To: talk@openstreetmap.org Subject: Re: [OSM-talk] A warning about gates and other barriers Message-ID: 20100920203358.743e0...@mum-quad Content-Type: text/plain; charset=UTF-8 So if you visit lots of places in ?outback? *Australia, you get to open **and close the gates as you go. **The gate should be closed, and you are free to pass, but have to open **the gate, pass the boundary and close the gate again.* You mean like in the movie The God's Must Be Crazy? :-) ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Richard Fairhurst schrieb: barrier=gate states that there's a gate. The thing about gates, as opposed to (say) walls, is that you can open them to get through. Not quite. It's that _someone_ can open them to get through. That someone does not have to be you, and you might not be able to get through after all. Robert Kaiser ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On Mon, Sep 20, 2010 at 3:12 PM, Donald Campbell II donaciano2...@gmail.com wrote: So if you visit lots of places in ?outback? Australia, you get to open and close the gates as you go. The gate should be closed, and you are free to pass, but have to open the gate, pass the boundary and close the gate again. You mean like in the movie The God's Must Be Crazy? :-) Same hemisphere, different continent, different desert. But the gates serve the same purpose, namely keeping the livestock fenced in. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Am 20.09.2010 13:48, Frederik Ramm: Hi, Claudius wrote: The point is that tags should only accommodate one piece of information. A gate is a gate, access is something else. Completely true. If a way has restricted access (and that may be physically enforced by the used of a closed gate) I tag the access restriction on the way behind the gate and not the gate. I don't see any added benefit by adding the access to a way behind a gate to the gate (!) itself. We both agree on the a gate is a gate, but a gate does not necessarily mean that a way has restricted access. Firstly, what is behind the gate differs depending on your location. Secondly, the way behind the gate may well be reachable by other means (i.e. a detour) - it is easy to imagine a gate where vehicles cannot pass, but still vehicles are allowed on both sides of the gate! Ah, true. Sorry for my rant only half thought through. Indeed access tagging on a gate makes sense :) And just to get my vote in as well: I used a plain barrier=gate in the sense that there's a gate, but I don't know about any access restrictions it may imply. Claudius ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 20.09.2010 13:20, Dave F. wrote: On 20/09/2010 10:31, Peter Wendorff wrote: Here we come to a great possibility for software to contribute to the OSM data. Wherever possible the software should provide a mechanism to add the data, if needed. How? The original point of this thread was that routing software couldn't distinguish it. Exactly - but IF a software navigates the user through e.g. a closed gate, the user can say this gate is closed for me as a cyclist. The software has to calculate a new route - and if that's recognized by the software, there could be a question back: why do I have to recalculate? Dependent on the current setting that could be reduced to specific map entities: - is the street missing? - is the gate closed and not passable for me? - is the street typed incorrectly (oneway, ...) - is a barrier missing (and what type of barrier) I think, there are a lot of small questions we can create to situations like the one in front of a gate, and these small, predefined questions/answers can hopefully contribute to the database in place, with end users as source without much skills using complex editors etc. You are right: The software couldn't distinguish it, but the user can - and with a cup of brain used for a feature like that the software can ask useful questions easy to answer by the user. In conclusion: The software can motivate and enable the user to distinguish it. regards Peter ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
2010/9/20 Dave F. dave...@madasafish.com: On 20/09/2010 09:07, M∡rtin Koppenhoefer wrote: I prefer asuming that a gate is closed (I tag them with access=private in these cases anyway) This is a bad assumption to have. Just because a gate is closed, it doesn't mean your not allowed access. maybe I wasn't clear: I am tagging gates with the access-restrictions that apply. I do this with access for closed gates (access=private/destination) and with exceptions like foot=yes, bicycle=yes, motorcar=yes. If applies, also add maxheight or height tags. I didn't care for motorcycles, wheelchairs, horses and others until now, but sometimes information for those can be taken from the kind of obstacle (gates exist in great variety why at least width is an important extra information, but others like [1] are more predictable for certain modes of movement) and I think we could also have different tags for pysically impossible allowed/forbidden. The first can sometimes be expressed by width and height (and maxweight, etc.), but not in all cases. The latter is what we usually describe with access and subtags (foot etc.). Another approach [2] to map physical possibility would be to define all possible barrier types and every application can decide based on the type and maybe measurements if it lets it's user through. barrier=gate means it's a gate nothing else. It could be open/closed all sorts of traffic could legitimately have access. yes Tag what you see on the ground. If you don't know, don't tag it. what also means: no tag - no information - this is opposed to the idea of defaults (which would be deriving the information from the fact that a tag is _not_ set). I agree. Cheers, Martin [1] http://wiki.openstreetmap.org/wiki/File:TR-a.JPG [2] http://wiki.openstreetmap.org/wiki/Proposed_features/New_barrier_types#Tag_values ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
[OSM-talk] A warning about gates and other barriers
Hello, I see a lot of frustration with routing not working when there are gates involved. In particular when the gate is in the middle of a public road (residential). This is because a gate with no access tags implies that nothing can go through. By contrast, most highway types let some type of traffic through. For routing purposes. it often creates islands, where some segments can't be reached from the rest of the routing graph. For example, much of this service road is cut off. http://www.openstreetmap.org/?lat=49.2214151lon=18.7401969zoom=18 Is there perhaps a worldwide tool that can identify these segments ? I tried keepright and the cloudmade debug map, but neither of them show it. Here is a more extensive description of the tagging of restricted areas: http://wiki.openstreetmap.org/wiki/Osm.org_Routing_Demo#How_to_tag_restricted_areas Regards, Nic ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 19/09/2010 14:37, Nic Roets wrote: This is because a gate with no access tags implies that nothing can go through. It wouldn't to me - no access tags on the gate would imply to me that nothing had been recorded about whether it was normally open or closed, or locked so that it couldn't be opened. If everyone else has the same understanding as you, perhaps someone needs to update http://wiki.openstreetmap.org/wiki/Tag:barrier%3Dgate so that it doesn't say ...which can be opened to allow or restrict access? Cheers, Andy ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 19/09/2010 14:37, Nic Roets wrote: For example, much of this service road is cut off. http://www.openstreetmap.org/?lat=49.2214151lon=18.7401969zoom=18 Part of the problem in this example, is that barrier=gate is tagged on the intersection. So its not clear whether the gate is across the service road or the pedestrianised road, or both. Best to avoid tagging gates on highway intersections. Maybe there could be a tool to check for this, eg Keepright? ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
2010/9/19 SomeoneElse li...@mail.atownsend.org.uk: On 19/09/2010 14:37, Nic Roets wrote: This is because a gate with no access tags implies that nothing can go through. It wouldn't to me - no access tags on the gate would imply to me that nothing had been recorded I agree with Andy here: no tags = no information. Which default is then implemented in a routing application depends on this application. To your question about a special tool: how would you know automatically which access restrictions apply? The example you pasted seems like bad mapping to me: a gate can't actually be directly on a bifurcation node. I would strongly recommend to tag it at it's real position, and if it's two gates (what would be a possible interpretation of your quoted example) then tag 2 separate gates. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Nic Roets wrote: This is because a gate with no access tags implies that nothing can go through. Where on earth do you get that idea from? barrier=gate states that there's a gate. The thing about gates, as opposed to (say) walls, is that you can open them to get through. Here are some pictures if I haven't explained it clearly enough: http://www.artlondon.com/photogallery/images/wellmann/Open-gate.jpg http://www.camulos.com/Virtual/wall.jpg cheers Richard -- View this message in context: http://gis.638310.n2.nabble.com/A-warning-about-gates-and-other-barriers-tp5547675p5547756.html Sent from the General Discussion mailing list archive at Nabble.com. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
2010/9/19 Richard Fairhurst rich...@systemed.net: barrier=gate states that there's a gate. The thing about gates, as opposed to (say) walls, is that you can open them to get through. Here are some pictures if I haven't explained it clearly enough: http://www.artlondon.com/photogallery/images/wellmann/Open-gate.jpg yes, I agree, but unless you tag it you also don't know if you really can open the gate. Here a picture for illustrational purpose: http://www.kpao.org/blog/assets_c/2008/12/locked-gate-thumb-400x266-155.jpg cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On Sun, Sep 19, 2010 at 10:33 AM, Richard Fairhurst rich...@systemed.net wrote: barrier=gate states that there's a gate. Doesn't it also state that there's a barrier? ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On Sun, Sep 19, 2010 at 11:16 AM, Anthony o...@inbox.org wrote: On Sun, Sep 19, 2010 at 10:33 AM, Richard Fairhurst rich...@systemed.net wrote: barrier=gate states that there's a gate. Doesn't it also state that there's a barrier? Nevermind. I see it is listed under access nodes and not barrier nodes. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
2010/9/19 Anthony o...@inbox.org: On Sun, Sep 19, 2010 at 10:33 AM, Richard Fairhurst rich...@systemed.net wrote: barrier=gate states that there's a gate. Doesn't it also state that there's a barrier? well, the simple key barrier doesn't actually state that there is a barrier but that there is something barrier-related. Barrier also contains values like gate, entrance which are indicating _openings_ in barriers. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
I see a lot of frustration with routing not working when there are gates involved. In particular when the gate is in the middle of a public road (residential). This is because a gate with no access tags implies that nothing can go through. This question has been in the back of my mind, for example, JOSM has a set of Allowed traffic tags. Does this mean the allowed traffic while the gate is closed, or open?Does this mean that a pedestrian can open the gate and freely pass at all times? Does access=yes imply opening times, or just mean that access is allowed as freely as the roadway access? ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On 19.09.2010 17:41, M∡rtin Koppenhoefer wrote: 2010/9/19 Anthonyo...@inbox.org: On Sun, Sep 19, 2010 at 10:33 AM, Richard Fairhurst rich...@systemed.net wrote: barrier=gate states that there's a gate. Doesn't it also state that there's a barrier? well, the simple key barrier doesn't actually state that there is a barrier but that there is something barrier-related. Barrier also contains values like gate, entrance which are indicating _openings_ in barriers. additionally most barriers have to be interpreted different for different use cases: A cycle barrier is no barrier for pedestrians, but for cyclists and car drivers. A bollard is a barrier for cars, but not for pedestrians and cyclists. A gate can be different - and therefore it should be tagged explicit using the access-tags. The barrier-tag in general should not be interpreted equal regardless of the value. Peter ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Nic Roets wrote: Nic Roets wrote: This is because a gate with no access tags implies that nothing can go through. Where on earth do you get that idea from? http://wiki.openstreetmap.org/wiki/Key:barrier in the sidebar under 'implies' And AFAIK that rule goes back to 2008. Wow. The OSM wiki never ceases to disappoint with its limitless provision of confusing, badly written, half thought-out crap. So we have a page that says implies access=no and then happily contradicts itself by saying an entrance that can be opened or closed to get through the barrier. That's can be opened (access=yes), not can't be opened (access=no). It's probably just as well the wiki documents, not defines. And given the vast preponderance of highway=gate nodes within (say) highway=footway ways, the wiki docs look pretty unambiguously wrong. Anyway, follow-ups to tagg...@. cheers Richard -- View this message in context: http://gis.638310.n2.nabble.com/A-warning-about-gates-and-other-barriers-tp5547675p5547906.html Sent from the General Discussion mailing list archive at Nabble.com. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On Sunday 19 September 2010 16:33:07 Richard Fairhurst wrote: Nic Roets wrote: This is because a gate with no access tags implies that nothing can go through. Where on earth do you get that idea from? barrier=gate states that there's a gate. The thing about gates, as opposed to (say) walls, is that you can open them to get through. It's a barrier. Most of the time it's there to keep people out and let only some people through. I always tag gates with the assumption that I don't have to tag who they keep out, but only who they let through. -- m.v.g., Cartinus ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
Hi, Cartinus wrote: I always tag gates with the assumption that I don't have to tag who they keep out, but only who they let through. Fair enough, but still I think it is far-fetched to assume that an un-tagged gate will be closed to all. When I map and I see a gate that is open but of which I cannot determine if, or when, it might not be so, I still tag it as a gate. I don't have more information - just that there is a gate and that it was open to traffic on a certain date and time. Now a conservative routing engine could of course say: If there's a gate and we don't know whether it'll be open, we assume it is closed. That's ok if the user is made aware of that, and it certainly is a decision for the routing engine to make. You cannot make the mapper responsible for deciding whether or not a car, or bicycle, routing can go through a barrier; he might not know. Treating all un-tagged gates as access=no will lead to mappers entering bogus access tags (or not adding the gate at all) in such cases. This is hardly desirable. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On Sun, Sep 19, 2010 at 6:02 PM, Richard Fairhurst rich...@systemed.net wrote: It's probably just as well the wiki documents, not defines. And given the vast preponderance of highway=gate nodes within (say) highway=footway ways, the wiki docs look pretty unambiguously wrong. So to obtain a definition of barrier=gate, you want me to do the following: Look at a region where barrier=gate has frequently been used without access tag. Let's say somewhere in the UK (a country I have never visited). Then conclude that a default of access=yes or foot=yes makes the most sense, because it will allow routing along some footways. Except there are places where the placements of the gates (the topology) would lead an intelligent person to conclude that the default access value for gates should be no. For example when there are many ways with gates into an area and one way without a gate: http://www.openstreetmap.org/browse/node/308935502 Then there is the possibility that a footway or gate have not been mapped. So I will have to look at many, many gates and have an outstanding intelligence to obtain the correct definition. I'd much rather just look at the wiki. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A warning about gates and other barriers
On Mon, Sep 20, 2010 at 4:54 AM, Frederik Ramm frede...@remote.org wrote: You cannot make the mapper responsible for deciding whether or not a car, or bicycle, routing can go through a barrier; he might not know. Treating all un-tagged gates as access=no will lead to mappers entering bogus access tags (or not adding the gate at all) in such cases. This is hardly desirable. The best thing OSM can do is to define how such cases are to be interpreted, and then for software using the data to follow those rules (or to break them for specific reasons). If you look at this thread, it's essentially here's how I read this, oh really? I do it differently, oh me too, but you know, everyone's opinion is valid! - which is pretty much useless. Personally, I think I prefer the default that access is open to anyone who could be on both sides. But anyway, could we try and reach consensus, and then document that? Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk