Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
On 28 April 2015 at 11:05, Colin Smale colin.sm...@xs4all.nl wrote: The existing through_route proposal may not be perfect but IMHO is a good base. It will need weeding through to keep it on-topic. This is how I see the scope of the discussion (just to get the ball rolling, feel free to shoot): 1) it has to be about junctions, not about individual ways (it's not about warning of sharp bends in a continuous road) 2) it has to be about aspects which cannot (reliably) be derived from the geometry alone (see point 1 above) 3) it must cover factors which affect the way the route to be driven is explained to the user (keep left vs. take the exit, follow the road to the right vs. turn right etc etc) 4) it *may* cover factors which affect the way the router chooses its optimum route (e.g. time penalties for a give way) //colin Is a 'through route' a continuation of - Road number - Road name - The lack of crossing white paint into/out of a 'side road' , or - Direction of travel I'm concerned about ambiguity arising out of potential different interpretations. I'm sure we will be able to find examples of all combinations of all of the above! -- Mike. @millomweb https://sites.google.com/site/millomweb/index/introduction - For all your info on Millom and South Copeland via *the area's premier website - * *currently unavailable due to ongoing harassment of me, my family, property pets* TCs https://sites.google.com/site/pmailkeey/e-mail ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
On 27/04/15 21:31, Rob Nickerson wrote: Ok a few people are agreeing that a relation is needed to assist the routing engine to provide higher quality instructions (with routing left unaffected). That's good. I'd like to get something in the wiki and ideally get it approved (this is not an invite to talk about the wiki or the approval process - I've heard it all before). Question: Should I revive the through_route proposal or start a new one under a different name, say route_continues (or just continues) so as to avoid any ambiguity with the use of through route in general language? Reading the objections on http://wiki.openstreetmap.org/wiki/Proposed_features/through_route and to be honest, the example used is simply wrong. Since AC and DC are both 'A56' then there is nothing additional needed. That routing software does not use the clean information in this case is not an OSM problem? Now if you want to tell the router that the 'main' route goes onto the other road that is a different problem? That the junction has a 'stop' when going from A to D is not unusual, with B-D having priority. That should be distinguishable if the correct junction tagging is applied? I was looking at the actual problem which only occurs when there is not a matching road segment either side of the junction. We do have the problem of not having reliable identification tags, but if we are working with road segments which have a usable ID, be that a 'ref' or second a road name, then this information is cleanly tagged on all three legs on the example. To my mind, the 'through_route' tag is also applied to those segments rather than the junction, so we have ... for example ... name=LondonRoad, through_route=OxfordWay and the reverse on the other side. The routing software then knows the primary route through the junction and can announce as appropriate. For unidentified roads, which is one of my irritations around this area, one needs a way of identifying the two ways which is not then also displayed. One of those places where 'object_id' could be useful, but to be honest I think one simply lives with the 'strange' messages. I don't think that things like trying to create a relation are productive. Although in the UK if the street gazetteer data becomes freely usable then there will be id's that can be used after reference and name. -- 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] Next: Relation name (WAS: Removing redundant routing instructions)
It's about that which cannot be inferred from the geometry and the current tagging. The road name and number come under current tagging, but sometimes they are not enough to know how to describe the next instruction(s) to the user. We are in the business of dictating the exact text that a navigator should speak, but we need to provide the basic information on which these decisions can be made. The ambiguity of a T-junction where the side-road has the same name and ref as one of the straight through legs is evident. Direction of travel is a good point - the through route may not be symmetrical. //colin On 2015-04-28 13:47, pmailkeey . wrote: On 28 April 2015 at 11:05, Colin Smale colin.sm...@xs4all.nl wrote: The existing through_route proposal may not be perfect but IMHO is a good base. It will need weeding through to keep it on-topic. This is how I see the scope of the discussion (just to get the ball rolling, feel free to shoot): 1) it has to be about junctions, not about individual ways (it's not about warning of sharp bends in a continuous road) 2) it has to be about aspects which cannot (reliably) be derived from the geometry alone (see point 1 above) 3) it must cover factors which affect the way the route to be driven is explained to the user (keep left vs. take the exit, follow the road to the right vs. turn right etc etc) 4) it *may* cover factors which affect the way the router chooses its optimum route (e.g. time penalties for a give way) //colin Is a 'through route' a continuation of * Road number * Road name * The lack of crossing white paint into/out of a 'side road' , or * Direction of travel I'm concerned about ambiguity arising out of potential different interpretations. I'm sure we will be able to find examples of all combinations of all of the above! -- Mike. @millomweb [2] - For all your info on Millom and South Copeland via THE AREA'S PREMIER WEBSITE - CURRENTLY UNAVAILABLE DUE TO ONGOING HARASSMENT OF ME, MY FAMILY, PROPERTY PETS TCs [3] ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk [1] Links: -- [1] https://lists.openstreetmap.org/listinfo/talk [2] https://sites.google.com/site/millomweb/index/introduction [3] https://sites.google.com/site/pmailkeey/e-mail ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
On 28/04/2015 13:23, Lester Caine wrote: Reading the objections on http://wiki.openstreetmap.org/wiki/Proposed_features/through_route and to be honest, the example used is simply wrong. How about this one: http://www.openstreetmap.org/node/21745867? Contrary to rational expectations, the 'through route' here is Groombridge Hill - Langton Road, and Ashurst Road has give-way lines. -- Steve --- This email has been checked for viruses by Avast antivirus software. http://www.avast.com ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
On 28 April 2015 at 08:21, Rob Nickerson rob.j.nicker...@gmail.com wrote: Please refer to my first email to the list (removing redundant routing instructions). Do you know how to view the mailing list archive online? Rob Yes thanks - I viewed archive stuff before posting - and in any case I've got the thread as e-mail ! I've looked at the case and your comments but feel the instruction is actually valid. There is a turn which is mentioned but likely the main purpose of the instruction is to let you know the change of road name. If you continue the route http://www.openstreetmap.org/directions?engine=osrm_carroute=52.45303%2C-1.48900%3B52.45434%2C-1.49086#map=18/52.45369/-1.48991layers=Q, you get the same message on a straight road but obviously without the 'turn' part. -- Mike. @millomweb https://sites.google.com/site/millomweb/index/introduction - For all your info on Millom and South Copeland via *the area's premier website - * *currently unavailable due to ongoing harassment of me, my family, property pets* TCs https://sites.google.com/site/pmailkeey/e-mail ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
Am 28.04.2015 um 13:47 schrieb pmailkeey . pmailk...@googlemail.com: Road number Road name The lack of crossing white paint into/out of a 'side road' , or Direction of travel I believe it's mainly about what appears to be road continuity (road markings). In Germany the traffic sign is this one: http://www.fahrschule-duenhoeft.de/cms/cms/old/zeichen/images/S202.JPG this is a scheme: http://www.mikrocontroller.net/attachment/157865/quiz3_03_12_720.jpg cheers Martin ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
On 28/04/15 15:15, Steve Doerr wrote: Reading the objections on http://wiki.openstreetmap.org/wiki/Proposed_features/through_route and to be honest, the example used is simply wrong. How about this one: http://www.openstreetmap.org/node/21745867? Contrary to rational expectations, the 'through route' here is Groombridge Hill - Langton Road, and Ashurst Road has give-way lines. It's not an unusual occurrence. Tagging Langton Road as through_route = Groombridge Hill would override the natural flow. Tag Ashurst Road as through_route = Langton Road gives that a higher priority than the 'A264' reference, and finally Groombridge Hill gets through_route = Langton Road as well. Although I could accept not tagging that since it is 'natural' once the A264 is ignored. http://wiki.openstreetmap.org/wiki/Talk:Proposed_features/through_route throws up a few more points. My point is that this is not a 'relation' problem, but rather that the through_route tag was getting mixed up with traffic management tagging. If you take out all the objections due to junction priority rules and properly document that area of tagging, one is left with only the cases that are missing a link through a junction? With the correct 'give way' tagged on the above junction, again the A264 'priority' is reduced but only if there is no matching restriction on Groombridge Hill? This does of cause create the need for a better micro-mapped junction layout since the highway=give_way tag needs to be a short distance down http://www.openstreetmap.org/way/285608726 and the routing software needs to understand the whole area of the junction rather than just the the common node. -- 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] Next: Relation name (WAS: Removing redundant routing instructions)
On 28 April 2015 at 13:15, Colin Smale colin.sm...@xs4all.nl wrote: It's about that which cannot be inferred from the geometry and the current tagging. The road name and number come under current tagging, but sometimes they are not enough to know how to describe the next instruction(s) to the user. We are in the business of dictating the exact text that a navigator should speak, but we need to provide the basic information on which these decisions can be made. The ambiguity of a T-junction where the side-road has the same name and ref as one of the straight through legs is evident. Direction of travel is a good point - the through route may not be symmetrical. //colin Are we then back down to the simple issue of failing to tag the arm(s) with the 'give way' markings ? - We're not tagging them in all necessary cases (where the 'side road' is in line with one arm of the main road) - Routers are ignoring the tags. Is this the solution rather than creating a new solution ? Is a 'through route' a continuation of - Road number - Road name - The lack of crossing white paint into/out of a 'side road' , or - Direction of travel I'm concerned about ambiguity arising out of potential different interpretations. I'm sure we will be able to find examples of all combinations of all of the above! -- Mike. @millomweb https://sites.google.com/site/millomweb/index/introduction - For all your info on Millom and South Copeland via *the area's premier website - * *currently unavailable due to ongoing harassment of me, my family, property pets* TCs https://sites.google.com/site/pmailkeey/e-mail ___ talk mailing listtalk@openstreetmap.orghttps://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk -- Mike. @millomweb https://sites.google.com/site/millomweb/index/introduction - For all your info on Millom and South Copeland via *the area's premier website - * *currently unavailable due to ongoing harassment of me, my family, property pets* TCs https://sites.google.com/site/pmailkeey/e-mail ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
On 28 April 2015 at 16:25, Colin Smale colin.sm...@xs4all.nl wrote: The give way sign won't help to distinguish between the arms where two roads diverge... You mean neither has give way markings ? Can you provide an example ? It's sounding a bit like Russia ! By the way, the sign is often a STOP sign, so the logic will have to check for both. Whether 'often' or not, that's true. That made me think of another peculiarity - but it's not a stop, just a give way on the through route. So yes, I've found a give way on the through route ! I guess this is a bit like the mini roundabout scenario. Can we accept the concept of not having a through route at a junction ? In the case I've found, there's no continuity - different number,different name, give way but is a continuation of the main route! -- Mike. @millomweb https://sites.google.com/site/millomweb/index/introduction - For all your info on Millom and South Copeland via *the area's premier website - * *currently unavailable due to ongoing harassment of me, my family, property pets* TCs https://sites.google.com/site/pmailkeey/e-mail ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
On 28 April 2015 at 16:05, Lester Caine les...@lsces.co.uk wrote: My point is that this is not a 'relation' problem, but rather that the through_route tag was getting mixed up with traffic management tagging. -- Lester Caine - G8HFL Is there need for a through route tag as well as the give way tag ? How is the whole concept affected by roundabouts (mini) - where all '3' arms are 'give way' What if the junction has more than one through route ? I'm thinking one arm in, one arm out (effectively dual carriageway) and one arm that's both in and out. -- Mike. @millomweb https://sites.google.com/site/millomweb/index/introduction - For all your info on Millom and South Copeland via *the area's premier website - * *currently unavailable due to ongoing harassment of me, my family, property pets* TCs https://sites.google.com/site/pmailkeey/e-mail ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
On 28/04/15 17:14, pmailkeey . wrote: My point is that this is not a 'relation' problem, but rather that the through_route tag was getting mixed up with traffic management tagging. Is there need for a through route tag as well as the give way tag ? Yes ... The give way/stop is traffic management and while one may be able to determine the named through route, just from that, it may not work both directions. How is the whole concept affected by roundabouts (mini) - where all '3' arms are 'give way' That is perhaps one of my biggest gripes with routing software which simply ignores them, at least a cross the mini-roundabout would help rather than 'turn right at roundabout' which is actually half a mile further on :( But this is where what is the 'straight on' route may well not be clear even on the ground, and so if the road id changes then some additional tag is necessary. What if the junction has more than one through route ? I'm thinking one arm in, one arm out (effectively dual carriageway) and one arm that's both in and out. Travel direction deals with the dual carriageway bit, but I can think of three mini-roundabout configurations I drive through regularly which OSMAND simply ignores, but have a total of three 'in' routes and one 'out' going one direction, but going the other direction is difficult to determine from the 'directions'. What is needed is proper lane directions through the junctions, rather than a 'left' or 'right' which are difficult even to decide when there is ALSO a road going across at 90degs :) 6 roads meet offset over two or three mini-roundabouts. -- 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] Next: Relation name (WAS: Removing redundant routing instructions)
Look at either of the examples in the through_route proposal page. If you are driving from bottom to top, you approach a Y-junction where two roads diverge. Whether there is a give-way sign coming in the opposite direction is not sufficient to understand which branch of the Y is the through route as seen from below as it were - the through route may not be symmetrical. I don't have an example of this to hand, but it wouldn't be right to exclude this possibility. And yes, there are junctions without a through route. In the USA you have four-way-stops whereby all roads see a stop sign. Sequencing is a bit informal - first-come-first-served in theory I believe, but that doesn't always work of course. Closer to home in continental Europe there are many junctions (away from main roads) which use priority from the right - i.e. all roads are deemed equal and no one road/route has priority over the others. These junctions are usually unmarked (i.e. no white lines and no signs) because they are deemed to be default in the absence of priority road signs (yellow diamonds). //colin On 2015-04-28 18:25, pmailkeey . wrote: On 28 April 2015 at 16:25, Colin Smale colin.sm...@xs4all.nl wrote: The give way sign won't help to distinguish between the arms where two roads diverge... You mean neither has give way markings ? Can you provide an example ? It's sounding a bit like Russia ! By the way, the sign is often a STOP sign, so the logic will have to check for both. Whether 'often' or not, that's true. That made me think of another peculiarity - but it's not a stop, just a give way on the through route. So yes, I've found a give way on the through route ! I guess this is a bit like the mini roundabout scenario. Can we accept the concept of not having a through route at a junction ? In the case I've found, there's no continuity - different number,different name, give way but is a continuation of the main route! -- Mike. @millomweb [2] - For all your info on Millom and South Copeland via THE AREA'S PREMIER WEBSITE - CURRENTLY UNAVAILABLE DUE TO ONGOING HARASSMENT OF ME, MY FAMILY, PROPERTY PETS TCs [3] ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk [1] Links: -- [1] https://lists.openstreetmap.org/listinfo/talk [2] https://sites.google.com/site/millomweb/index/introduction [3] https://sites.google.com/site/pmailkeey/e-mail ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
The existing through_route proposal may not be perfect but IMHO is a good base. It will need weeding through to keep it on-topic. This is how I see the scope of the discussion (just to get the ball rolling, feel free to shoot): 1) it has to be about junctions, not about individual ways (it's not about warning of sharp bends in a continuous road) 2) it has to be about aspects which cannot (reliably) be derived from the geometry alone (see point 1 above) 3) it must cover factors which affect the way the route to be driven is explained to the user (keep left vs. take the exit, follow the road to the right vs. turn right etc etc) 4) it *may* cover factors which affect the way the router chooses its optimum route (e.g. time penalties for a give way) //colin On 2015-04-27 22:31, Rob Nickerson wrote: Ok a few people are agreeing that a relation is needed to assist the routing engine to provide higher quality instructions (with routing left unaffected). That's good. I'd like to get something in the wiki and ideally get it approved (this is not an invite to talk about the wiki or the approval process - I've heard it all before). Question: Should I revive the through_route proposal or start a new one under a different name, say route_continues (or just continues) so as to avoid any ambiguity with the use of through route in general language? Cheers, Rob ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk [1] Links: -- [1] https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Next: Relation name (WAS: Removing redundant routing instructions)
The give way sign won't help to distinguish between the arms where two roads diverge... By the way, the sign is often a STOP sign, so the logic will have to check for both. //colin On 2015-04-28 17:09, pmailkeey . wrote: On 28 April 2015 at 13:15, Colin Smale colin.sm...@xs4all.nl wrote: It's about that which cannot be inferred from the geometry and the current tagging. The road name and number come under current tagging, but sometimes they are not enough to know how to describe the next instruction(s) to the user. We are in the business of dictating the exact text that a navigator should speak, but we need to provide the basic information on which these decisions can be made. The ambiguity of a T-junction where the side-road has the same name and ref as one of the straight through legs is evident. Direction of travel is a good point - the through route may not be symmetrical. //colin Are we then back down to the simple issue of failing to tag the arm(s) with the 'give way' markings ? * We're not tagging them in all necessary cases (where the 'side road' is in line with one arm of the main road) * Routers are ignoring the tags. Is this the solution rather than creating a new solution ? Is a 'through route' a continuation of * Road number * Road name * The lack of crossing white paint into/out of a 'side road' , or * Direction of travel I'm concerned about ambiguity arising out of potential different interpretations. I'm sure we will be able to find examples of all combinations of all of the above! -- Mike. @millomweb [2] - For all your info on Millom and South Copeland via THE AREA'S PREMIER WEBSITE - CURRENTLY UNAVAILABLE DUE TO ONGOING HARASSMENT OF ME, MY FAMILY, PROPERTY PETS TCs [3] ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk [1] ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk [1] -- Mike. @millomweb [2] - For all your info on Millom and South Copeland via THE AREA'S PREMIER WEBSITE - CURRENTLY UNAVAILABLE DUE TO ONGOING HARASSMENT OF ME, MY FAMILY, PROPERTY PETS TCs [3] ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk [1] Links: -- [1] https://lists.openstreetmap.org/listinfo/talk [2] https://sites.google.com/site/millomweb/index/introduction [3] https://sites.google.com/site/pmailkeey/e-mail ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
