Re: [OSM-talk] Cycle lanes & cycle tracks - my findings and a proposal
Thanks Martin, Yes, exactly right; I spent the time to do my research and also wrote up a brief introduction to cycle "tracks" vs "lanes". As suggested this has now been moved to the 'tagging' mailing list so feel free to follow the topic there. I'm a little concerned at how much the scope has widened since I started discussing cycle stuff, and am worried that some of the suggestions are so complex that the barrier to entry for newcomers would be set so high as to even put off current mappers. Fortunately common sense will prevail and we will end up with something better at the end of it - even just some clearer guidance on the current wiki page for the cycleway tag would certainly help me. :-) Rob ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Cycle lanes & cycle tracks - my findings and a proposal
Am 25.05.2012 um 10:20 schrieb Felix Hartmann : >> p.s. Personally I feel that cycle TRACKS would be much easier to map if >> drawn as a separate highway=cycleway >> > No, no,no,no > > If we want to change it, then we should > a) wait for the editors to support proper lane mapping He was writing about tracks not lanes > b) wait for the editors to support proper junction mapping > Proper junction mapping is mostly something the mapper has to do, less a question of the editor he uses (IMHO) > P.S. this is better placed on the tagging mailinglist +1 Cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Cycle lanes & cycle tracks - my findings and a proposal
p.s. Personally I feel that cycle TRACKS would be much easier to map if drawn as a separate highway=cycleway (despite any challenges the renderers and routers currently have with this) - it just makes things a lot easier!! No, no,no,no As for changing the cycleway key values: If we do that, we actually loose too much information. Yes, changing the current cycleway key, could be a rather good idea. But don't replace it by something not better. If we want to change it, then we should a) wait for the editors to support proper lane mapping b) wait for the editors to support proper junction mapping c) think about how to do it once the above is working and accepted. Don't change the cycleway key now. It ain't perfect, but with :left and :right it covers most situations and it isn't difficult to understand. Some places for thought: http://wiki.openstreetmap.org/wiki/Proposed_features/highway%3Djunction and http://wiki.openstreetmap.org/wiki/Proposed_features/Lane_group P.S. this is better placed on the tagging mailinglist ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Cycle lanes & cycle tracks - my findings and a proposal
2012/5/22 Janko Mihelić : > 2012/5/22 colliar >> The major problem I have with splitting cycleways of the highway is the >> missing reference to the highway. In Germany you have to use a cycleway >> by law (with some exceptions) and if the cycleway and the highway are >> mapped as two highways I do not get the information if the cycleway >> belongs to the road (have to use) or if it is a seperate road (can use). >> E.g. we need to find a solution how to connect all the objects of a road. > What if the cycleway belongs to a highway, but they are not very close, > there is a few meters of grass in between? I think connecting of any sort is > not necessary if you just use "bicycle=no" tag on the highway. -1, this has been discussed ad infinitum on the German list. Tagging the road with bicycle=no would be plain wrong, because it is not forbidden to use the road, you have the obligation to use the cycleway, which is quite different. E.g. if the cycleway is not suitable for your bike (e.g. you have a huge trailer, or it is damaged, or it is obstructed by stuff, or...) you will still be legally entitled to use the road. Also if the cycleway does not go where you want to go (say you wanted to turn left, and the cycleway doesn't permit this because of the kerb or some other obstacle) you can use the street. There is also other cases in which you can use the street despite there is a cycleway and the obligation to use it. cheers, Martin btw: shouldn't this thread take place in tagging? ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Cycle lanes & cycle tracks - my findings and a proposal
colliar wrote: On 22/05/12 12:13, Lester Caine wrote: > Martin Koppenhoefer wrote: >> 2012/5/21 Rob Nickerson: >>> p.s. Personally I feel that cycle TRACKS would be much easier to map if >>> drawn as a separate highway=cycleway (despite any challenges the >>> renderers >>> and routers currently have with this) - it just makes things a lot >>> easier!! >> >> +1, it is also more consistent and simple (seperate carriageway = >> distinct way). > > Personally I think we are reaching the point in a lot of areas where > representing a complex road as a single way simply because it's easier > for the renderers and routers is becoming a hindrance generally. Adding > tags for sidewalk, cycletrack and other details such as barriers between > carriageways is something that should just happen automatically from the > real mapped features? Not something that needs to be created manually > ignoring the features themselves ... The major problem I have with splitting cycleways of the highway is the missing reference to the highway. In Germany you have to use a cycleway by law (with some exceptions) and if the cycleway and the highway are mapped as two highways I do not get the information if the cycleway belongs to the road (have to use) or if it is a seperate road (can use). E.g. we need to find a solution how to connect all the objects of a road. THIS is exactly the problem that has yet to be addressed fully ... While some people have no interest in 'micro-mapping', the system does need to manage both layers of detail transparently. Viewing data at a larger zoom where the detail only needs to be provided as a tag on the base way - cycle, footpath and the like - while once zoomed in, the separate footpath, carriages, and other details are displayed as separate ways with the relevant detail such as grass verges and barriers. I'm not sure that 'relations' fill the gap here to relate micro-details to a single 'macro' way. A more integral mechanism would seem to be the right way to handle this? As a 'cycleway' is detailed, it is automatically linked with the base way that it is split from? Same with footpaths on either side of a road, where grass verges can be included as appropriate. -- 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// Firebird - http://www.firebirdsql.org/index.php ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Cycle lanes & cycle tracks - my findings and a proposal
2012/5/22 colliar > > The major problem I have with splitting cycleways of the highway is the > missing reference to the highway. In Germany you have to use a cycleway > by law (with some exceptions) and if the cycleway and the highway are > mapped as two highways I do not get the information if the cycleway > belongs to the road (have to use) or if it is a seperate road (can use). > E.g. we need to find a solution how to connect all the objects of a road. > > Cheers > Colliar > What if the cycleway belongs to a highway, but they are not very close, there is a few meters of grass in between? I think connecting of any sort is not necessary if you just use "bicycle=no" tag on the highway. Janko ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Cycle lanes & cycle tracks - my findings and a proposal
On 22/05/12 12:13, Lester Caine wrote: > Martin Koppenhoefer wrote: >> 2012/5/21 Rob Nickerson: >>> p.s. Personally I feel that cycle TRACKS would be much easier to map if >>> drawn as a separate highway=cycleway (despite any challenges the >>> renderers >>> and routers currently have with this) - it just makes things a lot >>> easier!! >> >> +1, it is also more consistent and simple (seperate carriageway = >> distinct way). > > Personally I think we are reaching the point in a lot of areas where > representing a complex road as a single way simply because it's easier > for the renderers and routers is becoming a hindrance generally. Adding > tags for sidewalk, cycletrack and other details such as barriers between > carriageways is something that should just happen automatically from the > real mapped features? Not something that needs to be created manually > ignoring the features themselves ... The major problem I have with splitting cycleways of the highway is the missing reference to the highway. In Germany you have to use a cycleway by law (with some exceptions) and if the cycleway and the highway are mapped as two highways I do not get the information if the cycleway belongs to the road (have to use) or if it is a seperate road (can use). E.g. we need to find a solution how to connect all the objects of a road. Cheers Colliar signature.asc Description: OpenPGP digital signature ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Cycle lanes & cycle tracks - my findings and a proposal
Martin Koppenhoefer wrote: 2012/5/21 Rob Nickerson: p.s. Personally I feel that cycle TRACKS would be much easier to map if drawn as a separate highway=cycleway (despite any challenges the renderers and routers currently have with this) - it just makes things a lot easier!! +1, it is also more consistent and simple (seperate carriageway = distinct way). Personally I think we are reaching the point in a lot of areas where representing a complex road as a single way simply because it's easier for the renderers and routers is becoming a hindrance generally. Adding tags for sidewalk, cycletrack and other details such as barriers between carriageways is something that should just happen automatically from the real mapped features? Not something that needs to be created manually ignoring the features themselves ... -- 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// Firebird - http://www.firebirdsql.org/index.php ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Cycle lanes & cycle tracks - my findings and a proposal
2012/5/21 Rob Nickerson : > p.s. Personally I feel that cycle TRACKS would be much easier to map if > drawn as a separate highway=cycleway (despite any challenges the renderers > and routers currently have with this) - it just makes things a lot easier!! +1, it is also more consistent and simple (seperate carriageway = distinct way). cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Cycle lanes
Excuse me while I but in... I'd agree that Left/right doesn't feel like the right solution - I've got a different idea for a solution which doesn't seem to have been suggested. For me the whole problem comes down to the fact that in the current representation there is no concept of a WIDTH of a way - or at least not one that is respected by any of the software. This means that positioning items at the 'edge' or 'inside' of a way becomes a complete mess. If instead we give the way a width then the problem becomes a lot simpler. The bus stop, cycle track or lane gets created in its actual physical location (as near as possible) and gets marked with a 'child of' relationship to the parent way. Width is either a standard default based on the 'highway' or an actual measurement. Because of the 'relationship' software that wants to render the way at a different size can either scale the way (and contents) or drop back to the standard way representation. Seems pretty clean to me - so what am I missing!? -- Brian On Wed, Apr 2, 2008 at 11:47 PM, Karl Newman <[EMAIL PROTECTED]> wrote: > On Wed, Apr 2, 2008 at 3:05 PM, Bjørn Bürger <[EMAIL PROTECTED]> > wrote: > > > > > Karl Newman wrote: > > > I don't know why everyone's opposed to left/right. It's unambiguous, > > > and properly structured it would not be difficult for > > > editors to accommodate it. > > > > Hmm, IMO neither north/south, nor left/right are a good solution for > > this problem. The only clean solution would be a relation, saying > > something like "feature=abc from node=x to node=y". > > > > Bjørn > > > > You still haven't solved the left/right problem. For example, house numbers > are commonly even on one side and odd on the other. How do you indicate > odd/even with the "from...to" structure you mention without using left/right > (or some equivalent)? Or what about bus stops which are only on one side of > the road (arguably a more difficult problem)? > > Karl > ___ > talk mailing list > talk@openstreetmap.org > http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk > > ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Karl Newman wrote: > You still haven't solved the left/right problem. For example, house > numbers are commonly even on one side and odd on the other. Not in Braunschweig: Many of our streets are numbered this way: 1 - 2 - 3 - 4 - 5 - [...] - 131 == \ [...] 136 - 135 - 134 - 133 - 133 > How do you indicate odd/even with the "from...to" > structure you mention without using left/right > (or some equivalent)? good point. Bjørn ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Wed, Apr 2, 2008 at 3:05 PM, Bjørn Bürger <[EMAIL PROTECTED]> wrote: > Karl Newman wrote: > > I don't know why everyone's opposed to left/right. It's unambiguous, > > and properly structured it would not be difficult for > > editors to accommodate it. > > Hmm, IMO neither north/south, nor left/right are a good solution for > this problem. The only clean solution would be a relation, saying > something like "feature=abc from node=x to node=y". > > Bjørn > You still haven't solved the left/right problem. For example, house numbers are commonly even on one side and odd on the other. How do you indicate odd/even with the "from...to" structure you mention without using left/right (or some equivalent)? Or what about bus stops which are only on one side of the road (arguably a more difficult problem)? Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Karl Newman wrote: > I don't know why everyone's opposed to left/right. It's unambiguous, > and properly structured it would not be difficult for > editors to accommodate it. Hmm, IMO neither north/south, nor left/right are a good solution for this problem. The only clean solution would be a relation, saying something like "feature=abc from node=x to node=y". Bjørn ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle Lanes
El Lunes, 31 de Marzo de 2008, Andy Allan escribió: > On Mon, Mar 31, 2008 at 3:59 PM, Peter Miller <[EMAIL PROTECTED]> wrote: > > Can I suggest that the vehicle oneway=yes/no attribute should be able to > > take an additional value of 'reverse' to make all the tags independent > > of the direction of the way and avoid the need to reverse ways at all. > > oneway = -1 has existed for quite some time, and is moderately widely used. > > yes| 199863 > 1 | 126642 > true | 99662 > no | 6348 > false | 1608 > -1 | 1359 > 0 |109 Speaking of which, I'll have to propose "oneway=reversible" - we have one piece of motorway that is one way on friday night, and the other way on sunday. Interestingly enough, this motorway tunnel does not appear on Gmaps, Ymaps, or even the IGN maps (our national mapping agency). Gotta use that point in my talks. -- -- Iván Sánchez Ortega <[EMAIL PROTECTED]> I reverently believe that the maker who made us all makes everything in New England, but the weather. I don't know who makes that, but I think it must be raw apprentices in the weather-clerks factory who experiment and learn how, in New England, for board and clothes, and then are promoted to make weather for countries that require a good article, and will take their custom elsewhere if they don't get it. -- Mark Twain signature.asc Description: This is a digitally signed message part. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle Lanes
On Mon, Mar 31, 2008 at 3:59 PM, Peter Miller <[EMAIL PROTECTED]> wrote: > > Can I suggest that the vehicle oneway=yes/no attribute should be able to > take an additional value of 'reverse' to make all the tags independent of > the direction of the way and avoid the need to reverse ways at all. I do > agree that attributes should use prefix values of forwards/backwards and > left/right and that editors should reverse these if the way is reversed. The very geeky "oneway => -1" has been around since the dawn of time. I think it's generally better to just reverse the way direction unless there's a specific reason why you can't. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle Lanes
I think Peter is spot on here. David On 31/03/2008 15:59, Peter Miller wrote: > Can I suggest that the vehicle oneway=yes/no attribute should be able to > take an additional value of 'reverse' to make all the tags independent of > the direction of the way and avoid the need to reverse ways at all. I do > agree that attributes should use prefix values of forwards/backwards and > left/right and that editors should reverse these if the way is reversed. > > Btw, this approach is used by GDF where a direction attribute is used which > can take the values 'traffic is allowed in both directions', 'traffic is > closed in the positive direction', 'traffic is closed in the negative > direction' and 'traffic is closed in both directions' (9.3.6). I believe > that this attribute in GDF can be used in conjunction with a vehicle class > to create different rules for different types of vehicle (including > pedestrians and cyclists). > > With regard to the debate about separate tracks or a 'handed' attributes for > the road I would suggest that there are times where either might be > appropriate, but that it would be reasonably easy to create a separate track > automatically from a 'handed attribute' if required (and would be much > easier than the reverse transformation) so the handed approach should be > preferred. > > I also suggest that it will be easier for the renderer to encode parallel > cycle lanes on cycle maps using the convention of colour coding the casing > of the road if the handed approach is used. Currently there are a lot of > problems with separate cycle tracks close to roads getting obscured by the > road itself or indeed obscuring the road. > > Personally I will continue to use handed attributes where the track is > parallel to the road and where there is no barrier between the track and the > road (other than some grass and a kerb). > > Fyi, the Cycle Data Standard for the Department for Transport in the UK that > I worked on in the autumn used the concept of an 'offset path' which had is > own identity (and could therefore be used in relationships) but which > borrowed its geometry from the main way to avoid all the problems of > stitching the way into all the side streets and to allow 'casing colour' > style maps to be created. I am requesting that they publish the standard so > we can compare and contrast and will let you know when it becomes available. > > > > > Regards, > > > > Peter Miller > > > > >> Message: 1 >> Date: Mon, 31 Mar 2008 11:41:26 + (UTC) >> From: David Dean <[EMAIL PROTECTED]> >> Subject: Re: [OSM-talk] Cycle lanes >> To: talk@openstreetmap.org >> Message-ID: <[EMAIL PROTECTED]> >> Content-Type: text/plain; charset=us-ascii >> >> Martin Vidner vidner.net> writes: >> >>> Make the prefixes "left:", "right:" special in the sense that when a >>> way is reversed, they get swapped. >>> So left:highway=bus_stop would become right:highway=bus_stop. >>> (Uh, maybe this is awkward for the renderer implementation. Could be >>> better to prefix the *value* instead: highway=left:bus_stop?) >> >> It seems to me that you could define the two sides of a way independent of >> the >> direction (if any) of the way. I'm just not sure what you would call the >> two >> sides. >> >> For example, lets start with "north" and "south". This would unambiguously >> define the two sides for all ways that are not running directly (or close >> to) >> north-south. "East" and "west" would work for those of course, but we want >> the >> same name no matter what the angle of the road. >> >> Maybe you could use "clockwise" and "anticlockwise" to define the side of >> that >> portion of the road you would get if you rotated it in that direction. >> >> So what I am basically getting at is that you don't need to define the >> side of >> the road based on the way direction, as it can be defined by the compass >> points, >> I'm just not sure what the two labels would be. Maybe "north-or-east" and >> "south-or-west" shortened to "noe" and "sow" could work if everything was >> clearly defined on the wiki. >> >> - David >> > > > > ___ > talk mailing list > talk@openstreetmap.org > http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk > ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle Lanes
On Mon, Mar 31, 2008 at 3:59 PM, Peter Miller <[EMAIL PROTECTED]> wrote: > > Can I suggest that the vehicle oneway=yes/no attribute should be able to > take an additional value of 'reverse' to make all the tags independent of > the direction of the way and avoid the need to reverse ways at all. oneway = -1 has existed for quite some time, and is moderately widely used. yes| 199863 1 | 126642 true | 99662 no | 6348 false | 1608 -1 | 1359 0 |109 Cheers, Andy ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Mon, Mar 31, 2008 at 12:41 PM, David Dean <[EMAIL PROTECTED]> wrote: > Martin Vidner vidner.net> writes: > > > Make the prefixes "left:", "right:" special in the sense that when a > > way is reversed, they get swapped. > > So left:highway=bus_stop would become right:highway=bus_stop. > > (Uh, maybe this is awkward for the renderer implementation. Could be > > better to prefix the *value* instead: highway=left:bus_stop?) > > > It seems to me that you could define the two sides of a way independent of > the > direction (if any) of the way. I'm just not sure what you would call the two > sides. > > For example, lets start with "north" and "south". This would unambiguously > define the two sides for all ways that are not running directly (or close to) > north-south. "East" and "west" would work for those of course, but we want > the > same name no matter what the angle of the road. > > Maybe you could use "clockwise" and "anticlockwise" to define the side of > that > portion of the road you would get if you rotated it in that direction. > > So what I am basically getting at is that you don't need to define the side > of > the road based on the way direction, as it can be defined by the compass > points, > I'm just not sure what the two labels would be. Maybe "north-or-east" and > "south-or-west" shortened to "noe" and "sow" could work if everything was > clearly defined on the wiki. > That doesn't work if your way goes round a corner. If I have a way which forms most of a loop its still ambiguous, and there's plenty more examples without getting that complicated. For most ways it's probably possible to pick an arbitrary direction where it's unambiguous, but you'll never cope with a roundabout, and I think most people will get very confused anyway, especially as editing the way's length may change the dominant line direction enough to flip the meaning. The only other way I can think of is to stick in a node on the correct side and link it with a relation that says "this side", but even that's susceptible to fairly simple editing mistakes. For proper "mistake proof" editing the only real solution is to use separate ways. So I'd say just leave direction as significant, but encourage use of separate ways instead of left/right tags or other equivalents. Dave ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Mon, Mar 31, 2008 at 6:41 AM, David Dean <[EMAIL PROTECTED]> wrote: > Martin Vidner vidner.net> writes: > > > Make the prefixes "left:", "right:" special in the sense that when a > > way is reversed, they get swapped. > > So left:highway=bus_stop would become right:highway=bus_stop. > > (Uh, maybe this is awkward for the renderer implementation. Could be > > better to prefix the *value* instead: highway=left:bus_stop?) > > > It seems to me that you could define the two sides of a way independent of > the > direction (if any) of the way. I'm just not sure what you would call the > two > sides. > > For example, lets start with "north" and "south". This would unambiguously > define the two sides for all ways that are not running directly (or close > to) > north-south. "East" and "west" would work for those of course, but we want > the > same name no matter what the angle of the road. > > Maybe you could use "clockwise" and "anticlockwise" to define the side of > that > portion of the road you would get if you rotated it in that direction. > > So what I am basically getting at is that you don't need to define the > side of > the road based on the way direction, as it can be defined by the compass > points, > I'm just not sure what the two labels would be. Maybe "north-or-east" and > "south-or-west" shortened to "noe" and "sow" could work if everything was > clearly defined on the wiki. > > - David > Sounds like you're assuming all streets are reasonably straight and/or progress in a well-defined direction. That may generally be the case over a long stretch of road, but it's not always the case (some roads tend to wind around a lot), especially if ways are split. I don't know why everyone's opposed to left/right. It's unambiguous, and properly structured it would not be difficult for editors to accommodate it. Reversing the direction of a way should be something that's done rarely anyway. Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Martin Vidner vidner.net> writes: > Make the prefixes "left:", "right:" special in the sense that when a > way is reversed, they get swapped. > So left:highway=bus_stop would become right:highway=bus_stop. > (Uh, maybe this is awkward for the renderer implementation. Could be > better to prefix the *value* instead: highway=left:bus_stop?) It seems to me that you could define the two sides of a way independent of the direction (if any) of the way. I'm just not sure what you would call the two sides. For example, lets start with "north" and "south". This would unambiguously define the two sides for all ways that are not running directly (or close to) north-south. "East" and "west" would work for those of course, but we want the same name no matter what the angle of the road. Maybe you could use "clockwise" and "anticlockwise" to define the side of that portion of the road you would get if you rotated it in that direction. So what I am basically getting at is that you don't need to define the side of the road based on the way direction, as it can be defined by the compass points, I'm just not sure what the two labels would be. Maybe "north-or-east" and "south-or-west" shortened to "noe" and "sow" could work if everything was clearly defined on the wiki. - David ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Mon, Mar 24, 2008 at 8:50 PM, Ben Laenen <[EMAIL PROTECTED]> wrote: > On Monday 24 March 2008, Alex Mauer wrote: > > > > It also has the problem that ways can easily get reversed, and then > > the left/right meanings are backwards. > > Then make editors change it automatically when reversing ways? Doesn't > look too difficult to me. > > Next to that we already have this kind of issue with routes with ways > with backward/forward roles anyway. Furthermore, cycle lanes will > always be part of the road, which can also have this left/right > problem, so I don't think this is a specific problem for cycle tracks. This is what I have been thinking about for a while (for the purpose of bus stops): Make the prefixes "left:", "right:" special in the sense that when a way is reversed, they get swapped. So left:highway=bus_stop would become right:highway=bus_stop. (Uh, maybe this is awkward for the renderer implementation. Could be better to prefix the *value* instead: highway=left:bus_stop?) For rendering, a bidirectional stop would be a full circle on a way/route and a one-way stop would be a half-circle. Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Andy Allan schrieb: > I'd tag every one of those as highway=cycleway on a separate way, if I > had the time and the patience. If I was busy, I would see > cycleway=track as being a stop-gap, and someone else could model them > as separate ways when they had the time (in the same way that there's > nothing wrong with a parking node when you can't/don't want to draw > the area in full). This is also my technique. Much too often those cycleways are very different from the motorcar lanes and require additional attention. For example, if time permits I'd like to add warning notes ("behind a bus_stop", "an entrance to a supermarket nearby", "beware of trash cans twice the week", "often use for parking car", etc.). > If, however, someone came along and removed my separate ways and added > cycleway=track to the road I would be very cross! BTW, I'd also like us to add cycleway=track to highway=cycleway way whenever the cycleway belongs to the street (in German "straßenbegleitend"). In Germany it is very important to know whether a cycleway belongs to the street, because then cyclists are force to use the cycleway. That's probably the same elsewhere in Europe. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Friday 28 March 2008, Andy Allan wrote: > I'd tag every one of those as highway=cycleway on a separate way, if > I had the time and the patience. If I was busy, I would see > cycleway=track as being a stop-gap, and someone else could model them > as separate ways when they had the time (in the same way that there's > nothing wrong with a parking node when you can't/don't want to draw > the area in full). > > If, however, someone came along and removed my separate ways and > added cycleway=track to the road I would be very cross! I've started thinking along those lines as well now. Separate highways for cycleways could provide some more detail. But keep this "lower resolution" method for quick tagging, so it doesn't need gps tracks for each cycleway before one is able to say in OSM that the road has a cycleway. However, if cycle tracks are separate ways, then cycle lanes like http://www.info.groenbrussel.be/images/random/fietspad.jpg should be as well IMHO (since the painted cycleways can just as well have all the extra tags as cycle tracks, they could just as well cross the street, be oneway or not, etc). The talk about legislation in the UK in one of the earlier responses to one of my mails makes me wonder though if their cycle lanes are somewhat the same as our "fietssuggestiestroken" (bicycle suggestion strips) like http://www.maldegem.be/infokrant_april07/0205.jpg The rules on it are quite simple: there are no rules for it :-). The strip is just part of the road, only in a different colour, but that has no meaning in traffic law whatsoever. Cars are allowed to drive on them whereas the one in the picture above cars can only cross them, and it doesn't give cyclists right of way over other traffic, whereas the first kind does. Doesn't take away the fact that renderers have to be able to handle it though before it gives nice results, and that may be a lot of work as well for all I can see. I also wonder about the exact interaction with sidewalks. Cycle tracks are often embedded into the sidewalk like this for example http://www.bttb.be/beelden/astridplein.JPG which is basically two footways on each side of the cycleway. How would one put such a situation in OSM? Greetings Ben ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Hi, > Interesting you should mention dual-carriageways -- there was some > discussion a while back: > > http://wiki.openstreetmap.org/index.php/Proposed_features/Left/right_things Thanks for bringing this to my attention, I have put a paragaph on the discussion page about why I strongly dislike direction-dependent tagging ;-) Bye Frederik ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Interesting you should mention dual-carriageways -- there was some discussion a while back: http://wiki.openstreetmap.org/index.php/Proposed_features/Left/right_things about how to "push things outwards" from road centrelines... On Tue, Mar 25, 2008 at 2:18 PM, Lars Aronsson <[EMAIL PROTECTED]> wrote: > OJ W wrote: > > > Sounds very similar to the cycleway tagging in Bedford; treat it > > as a separate way if it's not on the road, which makes it easy > > to show if it takes detours away from the road: > > But I want it to be just next to the street, with no gap and no > overlap, and getting this right requires that I know how it will > be rendered (which I don't), and there still is no chance I would > get it perfectly right. It does make more sense to draw one way > at the street's centerline and then use attributes to describe all > the various lanes for cars, buses, bikes and sidewalks as a list > with the width of each lane in metres. How far away is it that > Osmarender or Mapnik would render something like that? > > We have this problem with motorways already, that two ways > (northbound and southbound) run perfectly parallel in reality, but > on the map the distance between the ways is not always the same. > If this was mechanics, I would put a spacer washer between them. > > > -- > Lars Aronsson ([EMAIL PROTECTED]) > Aronsson Datateknik - http://aronsson.se > > ___ > talk mailing list > talk@openstreetmap.org > http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk > ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Mon, Mar 24, 2008 at 7:39 PM, Ben Laenen <[EMAIL PROTECTED]> wrote: > > OK, I've googled a bit for images of cycleways to get an idea about when > other people would tag a cycleway as a separate highway... (sorry, it's > a bit of a Belgium-centric selection...) [snip] Excellent stuff. I much prefer that these discussions are based on real life situations than hypotheticals. I'd tag every one of those as highway=cycleway on a separate way, if I had the time and the patience. If I was busy, I would see cycleway=track as being a stop-gap, and someone else could model them as separate ways when they had the time (in the same way that there's nothing wrong with a parking node when you can't/don't want to draw the area in full). If, however, someone came along and removed my separate ways and added cycleway=track to the road I would be very cross! Cheers, Andy ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Lars Aronsson wrote: | A bus stop is an attribute on a node (highway=bus_stop) in the | middle of a way. If I want to indicate that this bus stop is on | one side of the street, left and right don't matter much, since | there can be two ways both pointing towards the node. It would be | better to use the words north, east, west or south. If the way | goes from south-west to north-east, then "north" and "west" both | mean the left side of that way. I think it may be better to indicate the direction of the busses at the stop than the side of the road that the stop is on. So on a south-west, north-east road, both north and east would mean the busses travel in the direction of the way (and in the UK, the stop is on the left hand side of the way). J.D. Schmidt wrote: | It doesn't matter if the busstop is on the right or left side of the | road... Neither OSM wise, nor in the real world. Yes it does, because the side of the road determines which direction the busses go, which is a pretty important piece of information. Robert (Jamie) Munro -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (Darwin) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFH6ngaz+aYVHdncI0RAgEVAJ9/sOIKarF0j1umgi511nhjOFncUACeKedw POZACHq860ogfa+6qeCN9EA= =cZXj -END PGP SIGNATURE- ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Wed, Mar 26, 2008 at 10:39 AM, Jo <[EMAIL PROTECTED]> wrote: > Sebastian Spaeth wrote: > > Lars Aronsson wrote: > > > >> J.D. Schmidt wrote: > >> > >> > >>> It doesn't matter if the busstop is on the right or left side of > >>> the road... Neither OSM wise, nor in the real world. In the real > >>> world you use your eyes and see the busstop. > >>> > > > > > >> Of course it matters which side the bus stop is on. You don't > >> want to enter a bus that goes the wrong way, so you have to go to > >> the right stop. This is as fundamental as which bus lines the > >> stop is for. > >> > > > > I agree with LA here. Sometimes a bus stop would be on both sides of the > > road. Sometimes it would only be on one side while the one for the other > > direction is around the corner. Not having this kind of data is (in my > > personal data model view) not acceptable :-). > > Al left/right is a bit ambigious, so far I tend to draw little service > > roads and have a bus stop there. > > > > As for data, I tend to favor bus_line=123;direction=southampton or > > something similar. (or direction=north would also be a good way). > > > Over here, most bus stops serve for more than one line, so north/west > etc makes more sense to me than Brussels North Station/Airport personnel > parking lot. Which are totally different locations where the different > lines happen to go to. > > Polyglot > This seems to further highlight a need for a generic way to treat direction-sensitive information, in a way that can be processed by editors to automatically switch the left/right or whatever if a way is reversed. I know key namespaces are generally shouted down, but it would be a straightforward way to deal with this. i.e., left:bus_stop or left:house_number or left:my_heart_in_san_francisco. :-P If a way was reversed, the editors could change all the left: to right: Would need to apply to relations, too, though, and it starts to get tricky. But at a minimum, if a way is reversed, and it belongs to a relation containing direction-sensitive tags, the editors could prompt the user to confirm the action. Just brainstorming here, I welcome your thoughts. Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Sebastian Spaeth wrote: > Lars Aronsson wrote: > >> J.D. Schmidt wrote: >> >> >>> It doesn't matter if the busstop is on the right or left side of >>> the road... Neither OSM wise, nor in the real world. In the real >>> world you use your eyes and see the busstop. >>> > > >> Of course it matters which side the bus stop is on. You don't >> want to enter a bus that goes the wrong way, so you have to go to >> the right stop. This is as fundamental as which bus lines the >> stop is for. >> > > I agree with LA here. Sometimes a bus stop would be on both sides of the > road. Sometimes it would only be on one side while the one for the other > direction is around the corner. Not having this kind of data is (in my > personal data model view) not acceptable :-). > Al left/right is a bit ambigious, so far I tend to draw little service > roads and have a bus stop there. > > As for data, I tend to favor bus_line=123;direction=southampton or > something similar. (or direction=north would also be a good way). > Over here, most bus stops serve for more than one line, so north/west etc makes more sense to me than Brussels North Station/Airport personnel parking lot. Which are totally different locations where the different lines happen to go to. Polyglot ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Lars Aronsson wrote: > J.D. Schmidt wrote: > >> It doesn't matter if the busstop is on the right or left side of >> the road... Neither OSM wise, nor in the real world. In the real >> world you use your eyes and see the busstop. > Of course it matters which side the bus stop is on. You don't > want to enter a bus that goes the wrong way, so you have to go to > the right stop. This is as fundamental as which bus lines the > stop is for. I agree with LA here. Sometimes a bus stop would be on both sides of the road. Sometimes it would only be on one side while the one for the other direction is around the corner. Not having this kind of data is (in my personal data model view) not acceptable :-). Al left/right is a bit ambigious, so far I tend to draw little service roads and have a bus stop there. As for data, I tend to favor bus_line=123;direction=southampton or something similar. (or direction=north would also be a good way). ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
J.D. Schmidt wrote: > It doesn't matter if the busstop is on the right or left side of > the road... Neither OSM wise, nor in the real world. In the real > world you use your eyes and see the busstop. The same goes for oneway streets then. We don't have to indicate which direction they go, because you can see that with your own eyes once you get there. Eh? Of course it matters which side the bus stop is on. You don't want to enter a bus that goes the wrong way, so you have to go to the right stop. This is as fundamental as which bus lines the stop is for. -- Lars Aronsson ([EMAIL PROTECTED]) Aronsson Datateknik - http://aronsson.se ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
J.D. Schmidt wrote: > Lars Aronsson skrev: > >> Alex Mauer wrote: >> >> >>> It also has the problem that ways can easily get reversed, and >>> then the left/right meanings are backwards. >>> >> A bus stop is an attribute on a node (highway=bus_stop) in the >> middle of a way. If I want to indicate that this bus stop is on >> one side of the street, left and right don't matter much, since >> there can be two ways both pointing towards the node. It would be >> better to use the words north, east, west or south. If the way >> goes from south-west to north-east, then "north" and "west" both >> mean the left side of that way. >> >> >> > > Shall we paint the outhouse red, green or blue ?? > > It doesn't matter if the busstop is on the right or left side of the > road... Neither OSM wise, nor in the real world. In the real world you > use your eyes and see the busstop. With regards to OSM, you apply a tag > to the node indicating position of the busstop, relative to the road. > I.E : k="placement" v="left|right|middle" > > It doesn't even need to be rendered on the left or right of the road on > the map, but it could be. > > We are not trying to make a virtual copy of the world with the rendered > maps, we are trying to make a fairly accurate visual representation of > the world with the maps, that can be used to orientate and possibly even > navigate around in the real world. > > Compromises are made everyday on the renderings of both our and any > other maps. Thats one of the parts of cartography - Deciding what > compromises to use in order to make a map for a specific purpose. > > So just tag it with common sense, instead of having 2 months of > discussion on the mailing list - Next people will want to set down > subcommittees that need to deliver reports on best practices for mapping > the wastebaskets and their type at the busstops, which then has to go > through a hearing in the OSMF, and put to a consensus vote twice in a > year, before being vetoed by some goat-hearder mapping trails in outer > mongolia, since he doesn't think a wastebasket position should be mapped > at all, if goat-dung positions isn't mapped as well... > > We could call it "Working the EU-OSM way..." > At this moment in time I'm tagging bus stops as separate nodes on the side of ways. Given that the tag is highway=bus_stop I realise this is not correct, but I feel it's important to know on which side of the road that particular stop is. I fill out the ref numbers and of course it matters in which direction a certain bus will use that stop. So I'm glad we're having this discussion and I'm awaiting some consensus, so I can map the bus stops I find correctly from now on. Polyglot ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Lars Aronsson skrev: > Alex Mauer wrote: > >> It also has the problem that ways can easily get reversed, and >> then the left/right meanings are backwards. > > A bus stop is an attribute on a node (highway=bus_stop) in the > middle of a way. If I want to indicate that this bus stop is on > one side of the street, left and right don't matter much, since > there can be two ways both pointing towards the node. It would be > better to use the words north, east, west or south. If the way > goes from south-west to north-east, then "north" and "west" both > mean the left side of that way. > > Shall we paint the outhouse red, green or blue ?? It doesn't matter if the busstop is on the right or left side of the road... Neither OSM wise, nor in the real world. In the real world you use your eyes and see the busstop. With regards to OSM, you apply a tag to the node indicating position of the busstop, relative to the road. I.E : k="placement" v="left|right|middle" It doesn't even need to be rendered on the left or right of the road on the map, but it could be. We are not trying to make a virtual copy of the world with the rendered maps, we are trying to make a fairly accurate visual representation of the world with the maps, that can be used to orientate and possibly even navigate around in the real world. Compromises are made everyday on the renderings of both our and any other maps. Thats one of the parts of cartography - Deciding what compromises to use in order to make a map for a specific purpose. So just tag it with common sense, instead of having 2 months of discussion on the mailing list - Next people will want to set down subcommittees that need to deliver reports on best practices for mapping the wastebaskets and their type at the busstops, which then has to go through a hearing in the OSMF, and put to a consensus vote twice in a year, before being vetoed by some goat-hearder mapping trails in outer mongolia, since he doesn't think a wastebasket position should be mapped at all, if goat-dung positions isn't mapped as well... We could call it "Working the EU-OSM way..." Dutch ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Alex Mauer wrote: > It also has the problem that ways can easily get reversed, and > then the left/right meanings are backwards. A bus stop is an attribute on a node (highway=bus_stop) in the middle of a way. If I want to indicate that this bus stop is on one side of the street, left and right don't matter much, since there can be two ways both pointing towards the node. It would be better to use the words north, east, west or south. If the way goes from south-west to north-east, then "north" and "west" both mean the left side of that way. -- Lars Aronsson ([EMAIL PROTECTED]) Aronsson Datateknik - http://aronsson.se ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Lars Aronsson wrote: >Sent: 25 March 2008 2:19 PM >To: talk@openstreetmap.org >Subject: Re: [OSM-talk] Cycle lanes > >OJ W wrote: > >> Sounds very similar to the cycleway tagging in Bedford; treat it >> as a separate way if it's not on the road, which makes it easy >> to show if it takes detours away from the road: > >But I want it to be just next to the street, with no gap and no >overlap, and getting this right requires that I know how it will >be rendered (which I don't), and there still is no chance I would >get it perfectly right. It does make more sense to draw one way >at the street's centerline and then use attributes to describe all >the various lanes for cars, buses, bikes and sidewalks as a list >with the width of each lane in metres. How far away is it that >Osmarender or Mapnik would render something like that? > >We have this problem with motorways already, that two ways >(northbound and southbound) run perfectly parallel in reality, but >on the map the distance between the ways is not always the same. >If this was mechanics, I would put a spacer washer between them. > Most of the UK motorway network is not strictly parallel. In some places there is bigger divergence than a metre or two, elsewhere it varies because the width of the central reservation varies, especially to accommodate sight-lines around tighter curves. So, while I agree in general that probably the majority of the network is effectively parallel its not always the same distance from centre to centre of each carriageway. Cheers Andy > >-- > Lars Aronsson ([EMAIL PROTECTED]) > Aronsson Datateknik - http://aronsson.se > >___ >talk mailing list >talk@openstreetmap.org >http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
OJ W wrote: > Sounds very similar to the cycleway tagging in Bedford; treat it > as a separate way if it's not on the road, which makes it easy > to show if it takes detours away from the road: But I want it to be just next to the street, with no gap and no overlap, and getting this right requires that I know how it will be rendered (which I don't), and there still is no chance I would get it perfectly right. It does make more sense to draw one way at the street's centerline and then use attributes to describe all the various lanes for cars, buses, bikes and sidewalks as a list with the width of each lane in metres. How far away is it that Osmarender or Mapnik would render something like that? We have this problem with motorways already, that two ways (northbound and southbound) run perfectly parallel in reality, but on the map the distance between the ways is not always the same. If this was mechanics, I would put a spacer washer between them. -- Lars Aronsson ([EMAIL PROTECTED]) Aronsson Datateknik - http://aronsson.se ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On 25/03/2008, Sven Grüner <[EMAIL PROTECTED]> wrote: > > Bjørn Bürger schrieb: > > > Yes, but this is also the reality for cyclists. Everything involving > > cycleways is actually a mess, unfortunately. That is, because a > > bicycle is (mostly) not seen as an equal means of transportation. > > > Being considered a fanatical biker by my friends as well I share that > believe. > > > > But on the map, each distinct lane/track of a cycleway should > > be handled like e.g. the single lanes of a motorway/highway: Even the > > tiniest cycle-lane beneath a street has a different usage profile, > > different size and surface, different access rules, different right > > of way, etc. than the street. So IMO it clearly needs it's own way. > > > I can follow that argumentation based on the fact that dual carriageways > get separate ways. But in my opinion those are just as unfortunate. > Every traffic infrastructure (ie. road) consists of certain features > which IMHO should be represented by one single object in our DB holding > information about the features it's made from. I consider it really > strange that we currently map two roads instead of one only because the > real road has the feature "hard shoulder in the middle". It would make sense to only need to tag a road once, no matter how many parts it's made of, hopefully one day relationships will be handled well enough by editors and renderers that that is what happens... If a road has multiple parts to it, i.e. pavement, cycleways, multiple lanes in different directions for different uses etc, you want to keep them all linked together, tagged once, but, having separate sub-objects describing those features has a use... You will probably need multiple relations though and doing this is going to make the editors and renders more complex for sure, even if it's just the code that gets more complex to try and keep the interface simple... Renderers can of course be trimmed to compensate all the disadvantages > that come with this strategy or mappers can be encouraged to use > relations to glue these together again but that's not really solving the > problem, but creating workarounds for every purpose it encounters. You could split the data into two or more datasets, with one, where everything is recorded in as much details as possible as accurately as possible, maybe even splitting every single lane including turning lanes, and a second, where everything is recorded as single ways with each of those ways containing lots of tags to try and describe things but loosing lots of accuracy in the process... Advantages and disadvantages to both... The hurdles to overcome when it comes to rendering are probably pretty equal though... And think how much effort would be involved in making two openstreetmaps... To me, getting all the information possible recorded as accurately as possible would have to be the goal, then no matter what you want to use it for, the data is there, yep, it might be more difficult to work with than a simple dataset if you only have a simple use, but, if you want to do something more complex then not having the data makes it not possible... I believe in relationships as being the key to pulling complex data together and I believe in aiming for more detail than we can even have right now with current tagging... This worked fine when focussing on car-traffic but when we really want > to provide high-quility (usable for routing/navigation) data of footways > and cycleways I'm afraid we need a different approach. > > I'm not saying it's impossible by the way we do it now but I envision > there must be a better way... > > > > It will get easier, if support for that stuff is added to the editors. > > > There are people who already believe our editors are too complex (not me). I think there's definite room for relations being more polished/integrated into the editors... Right now, the core functionality is there, but it's not very usable for everyday mapping and tagging... I'm sure that's going to improve over time though... By the way, to follow the links to cycleways theme of the thread... http://www.informationfreeway.org/?lat=30.276479076977562&lon=120.16459191145113&zoom=17&layers=B000F000F This shows a junction of a dual carriageway with an elevated trunk dual carriageway... Sandwiched between them is a cycle flyover... This is one road that could benefit from more detail... to the west, the road is actually wider with a wide pavement, curb down to cycleway, barrier/garden, two lanes of traffic in one way, barrier/garden, bus lane with curbing on the side it meets the next lanes, two more lanes of traffic in one way, widening to 3 at points and all merging into 5-7 lanes at junctions with solid lines in the middle and the same layout repeated on the other side... From a navigation point of view, this is a nightmare, you can only turn from certain parts of the road, the barriers/gardens only have a few openings... d __
Re: [OSM-talk] Cycle lanes
Sven Grüner wrote: > Being considered a fanatical biker by my friends as well I share that > believe. ;-) > This worked fine when focussing on car-traffic but when we really want > to provide high-quility (usable for routing/navigation) data of footways > and cycleways I'm afraid we need a different approach. True. But it's hard to think of an abstraction, for something like ... "This is a highway with duty-to-use cycleways on either side, but the cycleway on the left side is mostly on a different level of height, has a paved surface and looses right of way on some crossings (because it slightly moves away from the main highway while there is no way to switch to the street before that). The other cycleway has an unpaved surface, is mostly oneway, except the part between somewhere and somewhereelse (and between ... and ... you are free to use either the street or the cycleway whereas on some parts the cycleway is incomplete or missing" (That was the description for my way to work and I have seen stuff like that all over Europe so far) > You're coming to tomorrows Stammtisch? yes. Bjørn ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Hi, > > For now, I would advocate extra ways whereever the cycleway is not > > just a lane painted on the road, and editors might become smart enough > > to detect a bordering cycleway and move that together with the road if > > you move the road or so. > > Please, avoid such "smartness"; some editor actions are already too > greedy. ;) Oh, don't worry, we'll use a modifier key for it ;-) Bye Frederik -- Frederik Ramm ## eMail [EMAIL PROTECTED] ## N49°00'09" E008°23'33" ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Bjørn Bürger schrieb: > Yes, but this is also the reality for cyclists. Everything involving > cycleways is actually a mess, unfortunately. That is, because a > bicycle is (mostly) not seen as an equal means of transportation. Being considered a fanatical biker by my friends as well I share that believe. > But on the map, each distinct lane/track of a cycleway should > be handled like e.g. the single lanes of a motorway/highway: Even the > tiniest cycle-lane beneath a street has a different usage profile, > different size and surface, different access rules, different right > of way, etc. than the street. So IMO it clearly needs it's own way. I can follow that argumentation based on the fact that dual carriageways get separate ways. But in my opinion those are just as unfortunate. Every traffic infrastructure (ie. road) consists of certain features which IMHO should be represented by one single object in our DB holding information about the features it's made from. I consider it really strange that we currently map two roads instead of one only because the real road has the feature "hard shoulder in the middle". Renderers can of course be trimmed to compensate all the disadvantages that come with this strategy or mappers can be encouraged to use relations to glue these together again but that's not really solving the problem, but creating workarounds for every purpose it encounters. This worked fine when focussing on car-traffic but when we really want to provide high-quility (usable for routing/navigation) data of footways and cycleways I'm afraid we need a different approach. I'm not saying it's impossible by the way we do it now but I envision there must be a better way... > It will get easier, if support for that stuff is added to the editors. There are people who already believe our editors are too complex (not me). > Bjørn You're coming to tomorrows Stammtisch? Grüße, Sven ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Ben Laenen wrote: > I beg to differ here. When you have to tag cycleways belonging to a road > not as "highway=whatever, cycleway=track" but as > separate "highway=cycleway" they just become an editing mess, > especially at intersections. Yes, but this is also the reality for cyclists. Everything involving cycleways is actually a mess, unfortunately. That is, because a bicycle is (mostly) not seen as an equal means of transportation. But on the map, each distinct lane/track of a cycleway should be handled like e.g. the single lanes of a motorway/highway: Even the tiniest cycle-lane beneath a street has a different usage profile, different size and surface, different access rules, different right of way, etc. than the street. So IMO it clearly needs it's own way. > Especially when adding route relations to > them. Just imagine two dual carriage ways with on either side a > cycleway crossing: you then need 24 different ways to just represent > that one intersection, like this: > > > ---- cycleway > ---- road > ---- road > ---- cycleway > Yes, that's odd. But it is reality and we have to live with it. It will get easier, if support for that stuff is added to the editors. > * Here's the big argument: There's no information lost by adding tags to > the road like "cycleway=track" (we need a few more tags though for the > more exotic cases, like when the cycle track also serves as sidewalk), Interesting idea - but then you should also skip all those different highway=* types in favour of a generic way. A big motorway could be added with something like ... highway=true highway_type=high_cypacity_motorway_for_fast_vehicles highway_carriageways=2 highway_carriageway_left_lanes=2 highway_carriageway_right_lanes=4 [...] wow, that's cute ;-)¹ > * You can usually arbitrarily go from the cycleway to the main road (to > cross it for example). Though this might be true for your personal reality, a osm based routing algorithm shouldn't do anything like that. No navigation system would advise you to do a u-turn on a big motorway (at any arbitrary location). Why should this be true for a cycle map? > I've tagged the cycleways as their own highway once, but just doing that > I got quickly convinced that doing that was just a bad idea You are right. At the moment, it is a bad idea to just map things like that. We have to think this through, maybe ask the routing people for advice and replace those myriads of different descriptions in the wiki with one description, which actually works for all [tm] ;-)² Bjørn ¹) You may want to add some virtual irony=more_or_less_sublte tag here ²) Yes, i know - nearly impossible. But we can try³... ³) I am just trying to get all the peaces together... ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
> For now, I would advocate extra ways whereever the cycleway is not > just a lane painted on the road, and editors might become smart enough > to detect a bordering cycleway and move that together with the road if > you move the road or so. Please, avoid such "smartness"; some editor actions are already too greedy. ;) ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
> Next to that we already have this kind of issue with routes with ways > with backward/forward roles anyway. Furthermore, cycle lanes will > always be part of the road, which can also have this left/right > problem, so I don't think this is a specific problem for cycle tracks. Why shall c/w suffer from the same limitation? Let's avoid this problem right from the beginning and just draw the c/w where it belongs. For sure, I do not want to store all the c/w characteristics within a single underscore connected attribute as subattributes. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Monday 24 March 2008, Dave Stubbs wrote: > Personally I'd start to way tracks separately when they have a clear > separation. That's deliberately ambiguous because I think it varies. > But yeah a 10m gap would certainly do it, but even a 1m gap if it's > made of something very solid. OK, I've googled a bit for images of cycleways to get an idea about when other people would tag a cycleway as a separate highway... (sorry, it's a bit of a Belgium-centric selection...) Cycleway separated by 30cm of grass, many connections at driveways: http://www.soundandvision.be/rijbewijs/wegwet/afbeeldingen/wegwet11a.JPG One meter of grass between cycleway and road, still quite a number of connections to the road: http://users.nucleus.be/adri/gfx/fietspad5.jpg Instead of grass, some small bushes: http://www.provant.be/binaries/aanleg%20subs%20westerlo%20gevaertlaan_tcm7-13420.jpg Not a cycle lane, but not separated by a kerb: http://skender.be/blog/afbeeldingen/VoskenslaanVerkeerslichtB.jpg Separated by a 50cm wide platform: http://www.laurencelibert.be/assets/1308/fietspad_article.jpg Separated by a line of trees, no kerb: http://www.mobielvlaanderen.be/figs/convenants/MOB8320F01.jpg Separated by 5m of grass: http://www.vldekeren.be/wp-images/media/puihoek1.jpg Parked cars, small patch of grass with some trees: http://blogsimages.skynet.be/images_v2/002/516/036/20070710/dyn009_original_640_480_pjpeg_2516036_7f35443c9fa0bdee2e8faf85200bc3ab.jpg Diagonally parked cars, some grass and trees: http://blogsimages.skynet.be/images_v2/002/516/036/20070710/dyn009_original_640_480_pjpeg_2516036_73826f8b3ac379fa17660f754ef96e9b.jpg My opinion is quite easy: they're all just "cycleway=track" as part of the main road, but I'd like to see what others do in each of these situations. Greetings Ben ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Hi, > highway=secondary > cycleway:left=bidirectional_track > cycleway:right=track > > highway=secondary > cycleway=left_bidirectional_track;right_track > > The first won't fly because everytime anybody mentions namespaces it gets > boo-ed away as being too complex. As you correctly say, both these solutions are horrible (especially if you think further and come to "cycleway:left:distance_from_road=5m" and all that). In an ideal world, I'd say let us have extra ways for everything, and some kind of linkage (relations or anything) between them. A renderer could then leave out road-adjacent cycleways if so desired, and an editor could move them together with the road they're linked to. For now, I would advocate extra ways whereever the cycleway is not just a lane painted on the road, and editors might become smart enough to detect a bordering cycleway and move that together with the road if you move the road or so. > But my main point stays that putting everything into that one way-object > makes > it much too hard to see what is tagged as what during the editing phase, > resulting in numerous errors. I think it would not be impossible to improve editors to the point where they could work with your above examples, but that would not reduce their horribleness. Bye Frederik -- Frederik Ramm ## eMail [EMAIL PROTECTED] ## N49°00'09" E008°23'33" ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Monday 24 March 2008, Alex Mauer wrote: > Cartinus wrote: > > The first won't fly because everytime anybody mentions namespaces > > it gets boo-ed away as being too complex. The second one is > > absolutely no fun to write stylesheets for (or the renderer needs a > > preprocessor to split them). Neither way of tagging is expandable > > if you want to put more properties on them (paved e.g.) > > It also has the problem that ways can easily get reversed, and then > the left/right meanings are backwards. Then make editors change it automatically when reversing ways? Doesn't look too difficult to me. Next to that we already have this kind of issue with routes with ways with backward/forward roles anyway. Furthermore, cycle lanes will always be part of the road, which can also have this left/right problem, so I don't think this is a specific problem for cycle tracks. Greetings Ben ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On 24/03/2008 15:12, Ben Laenen wrote: > On Monday 24 March 2008, Dave Stubbs wrote: >>> ---- cycleway >>> ---- road >>> ---- road >>> ---- cycleway >> I count 8 ways? >> Unless you are splitting all the ways at absolutely every >> intersection which is probably a little excessive. In Cambridge (where we probably have more than most in the UK) I took the view that where a cycleway is physically separate from the road I would mark it as a separate highway, in the same way that separate carriageways of a dual carriageway road are separate highways. They sometimes diverge away from the road they are parallel to and generally do different things to the road. Consider what's going on here, for example (the E-W route is part of NCN51, the N-S one is local): http://www.openstreetmap.org/?lat=52.20915&lon=0.18708&zoom=16&layers=B0FT (and look what happens at points further east, south and west). What I haven't done at all in Cambridge is marked cycle lanes, or shared use footways which are separate from the road only by a kerb. I do want to do this in due course, but we didn't have relations when I started Cambridge. In general the cycleways I've done are linked up properly at junctions. However a case I need to correct is where there is a T junction opposite, which is often like this: c/w --+- road | I should do the following, where true, or routing doesn't work properly (this is just like a road meeting a dual carriageway, where there may or may not be a gap in the central reservation). --+- c/w | --+- road | ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Cartinus wrote: > The first won't fly because everytime anybody mentions namespaces it gets > boo-ed away as being too complex. The second one is absolutely no fun to > write stylesheets for (or the renderer needs a preprocessor to split them). > Neither way of tagging is expandable if you want to put more properties on > them (paved e.g.) It also has the problem that ways can easily get reversed, and then the left/right meanings are backwards. -Alex Mauer "hawke" ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Mon, Mar 24, 2008 at 3:12 PM, Ben Laenen <[EMAIL PROTECTED]> wrote: > On Monday 24 March 2008, Dave Stubbs wrote: > > > ---- cycleway > > > ---- road > > > ---- road > > > ---- cycleway > > > > I count 8 ways? > > Unless you are splitting all the ways at absolutely every > > intersection which is probably a little excessive. > > Not if you need to have route relations going over them (bus routes, > cycle routes, walking routes etc). Then you often need access to each > separate piece. A person tends to think differently when you start > doing cycle routes on roads, that's how I stopped doing it, since I > couldn't see the point in spending hours of extra work in putting in > cycle tracks next to all roads and splitting up every intersection when > simple cycleway=* tags give exactly the same information (and given the > fact that yahoo imagery often couldn't give enough details about the > exact location anyway, not to mention the poor people who can only work > from GPS tracks who need to track the same road over and over again for > car/bicycle/pedestrian). It turned out that tagging a route with > highway=cycleway just took me three times more work, and made me do > numerous more mistakes. > > Sure, you can now just call me lazy for not doing that work, but note > that part about giving exactly the same information. If it's all that close I'd tend to agree. You're certainly not crazy :-) But if there's clear separation then the real question is what you do when someone comes along and "corrects" it. If it really does just follow the road then you are right about it containing the same information (topologically at least, but as you say the location of each is likely to be estimation anyway). > > > > Obviously creating a way for every single cycle lane is going to just > > cause a mess, so where they do just follow the road, on the road, > > it's probably best to keep them as just a simple tag. > > However, where they are clearly separate, it's probably best to tag > > them as you would a dual-carriageway, and for the same reasons. With > > a separate cycleway you generally can't just hop-on/hop-off without > > being a menace to other traffic and there's sometimes even a physical > > barrier; the way can also diverge from the main road way, taking > > short-cuts round roundabouts or similar. > > And shortcuts still get their own highways since that's a point where > they diverge from the main road. But for 99% of the cases where the > cycleway just follows the road that's overkill. Nobody is going to have > doubts about where the cycleway is exactly: next to the road. > > > > > But there are more reasons why I don't like these as separate > > > highways: > > > > > > * We're also not tagging sidewalks as separate "highway=footway" > > > right (well, I guess there is not tag for sidewalks yet but it'll > > > come -- but I can't imagine someone tagging them all like separate > > > ways anyway, just think about the intersection mentioned above and > > > add four "highway=footway"s to them). Cycleways are usually between > > > the sidewalk and the road, so it becomes quite odd that a sidewalk > > > is just a tag, but a cycleway is its own highway. > > > > That's just an argument for modeling the pavement properly. Most > > pavements are just tacked onto the road as an extra "lane", but with > > a kerb (to discourage the cars from using it :-) ), so I wouldn't > > usually bother adding these as separate ways, but where the pavement > > diverges or is clearly separate, it should probably be modelled as > > such. > > But where is the point between being separated from the road and not? As > a cyclist I can usually get off my bike and walk on the pavement next > to it at any point. I can also cross the road at any point. A line of > trees doesn't stop that. Some grass doesn't stop that, a line of parked > cars doesn't either. Of course, at the exact location of a tree or > perhaps a flower bed I can't hop from cycle lane to whatever, but I > can't see a pavement being tagged like lots of small highways next to > the road at every tree: > > /| > \| > | > /| > \| > > Anyway, the exact distinction between cycle lane and cycle track seems > quite odd to me: in legislation there's no difference between them > anyway, both give exactly the same rights/obligations as a road user > (at least here in Belgium). > > So I guess everyone agrees that a cycleway which is painted right next > to the part of the road where cars drive is a cycle lane. But at what > point does it become a track? > > Does a line of 10cm high and 50cm long concrete blocks make it a cycle > track? A 5m wide area painted diagonal stripes where no-one can drive > between cycleway and "motorcarway"? A kerb? Parking spaces? A 50cm wide > patch of grass? A line of trees? A 5m wide patch of grass? A flower > bed? > > So where's the difference? I te
Re: [OSM-talk] Cycle lanes
On Monday 24 March 2008 16:13:58 Dave Stubbs wrote: > On Mon, Mar 24, 2008 at 2:38 PM, Cartinus <[EMAIL PROTECTED]> wrote: > > On Monday 24 March 2008 15:18:00 Dave Stubbs wrote: > > > There are always edge cases that cause problems. > > > > None of the things I mentioned above are really rare here in the > > Netherlands. I realise this country is not very big, so maybe we have to > > live with the fact we are an "edge country"? > > Most of the things that you mentioned can be covered with the > cycleway=left_lane style of tagging. And with modifications to mapnik > it would be possible to render those, but there maybe some edge case > problems. For many things it may be better and cleaner to tag like > that. > > Sometimes it maybe better to create separate ways. > These are things we need to sort out. > > Please don't be so quick with the sarcasm, it's not very constructive. There are basically two options to try and put separate properties on the left- and right handed side cycleway while keeping everything on one way-object: highway=secondary cycleway:left=bidirectional_track cycleway:right=track highway=secondary cycleway=left_bidirectional_track;right_track The first won't fly because everytime anybody mentions namespaces it gets boo-ed away as being too complex. The second one is absolutely no fun to write stylesheets for (or the renderer needs a preprocessor to split them). Neither way of tagging is expandable if you want to put more properties on them (paved e.g.) While mapnik might be able to displace stuff sideways, last I heard this was impossible in XSLT/SVG But my main point stays that putting everything into that one way-object makes it much too hard to see what is tagged as what during the editing phase, resulting in numerous errors. -- m.v.g., Cartinus P.S. There was nothing meant as sarcastic in my previous post, only the first one in this thread was, because what Ben wrote is IMHO very for from the "truth". ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
>> On Mon, Mar 24, 2008 at 7:48 AM, Mike Collinson <[EMAIL PROTECTED]> wrote: >> > As a cyclist currently in Stockholm, I personally draw a separate way >> > and label it highway=cycleway, cycleway=track, foot=yes, >> surface=paved. > > I think the surface=paved is redundant too, because that is what I would > expect if no surface tag was put on the way. Because I do not the surface if I cannot remember me (surface=uncertain?), a little bit redundancy does not hurt. This I know for sure, what's the matter, esp. because in Germany many cycleways are unpaved... ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Mon, Mar 24, 2008 at 2:38 PM, Cartinus <[EMAIL PROTECTED]> wrote: > On Monday 24 March 2008 15:18:00 Dave Stubbs wrote: > > There are always edge cases that cause problems. > > None of the things I mentioned above are really rare here in the Netherlands. > I realise this country is not very big, so maybe we have to live with the > fact we are an "edge country"? Most of the things that you mentioned can be covered with the cycleway=left_lane style of tagging. And with modifications to mapnik it would be possible to render those, but there maybe some edge case problems. For many things it may be better and cleaner to tag like that. Sometimes it maybe better to create separate ways. These are things we need to sort out. Please don't be so quick with the sarcasm, it's not very constructive. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Monday 24 March 2008, Dave Stubbs wrote: > > ---- cycleway > > ---- road > > ---- road > > ---- cycleway > > I count 8 ways? > Unless you are splitting all the ways at absolutely every > intersection which is probably a little excessive. Not if you need to have route relations going over them (bus routes, cycle routes, walking routes etc). Then you often need access to each separate piece. A person tends to think differently when you start doing cycle routes on roads, that's how I stopped doing it, since I couldn't see the point in spending hours of extra work in putting in cycle tracks next to all roads and splitting up every intersection when simple cycleway=* tags give exactly the same information (and given the fact that yahoo imagery often couldn't give enough details about the exact location anyway, not to mention the poor people who can only work from GPS tracks who need to track the same road over and over again for car/bicycle/pedestrian). It turned out that tagging a route with highway=cycleway just took me three times more work, and made me do numerous more mistakes. Sure, you can now just call me lazy for not doing that work, but note that part about giving exactly the same information. > Obviously creating a way for every single cycle lane is going to just > cause a mess, so where they do just follow the road, on the road, > it's probably best to keep them as just a simple tag. > However, where they are clearly separate, it's probably best to tag > them as you would a dual-carriageway, and for the same reasons. With > a separate cycleway you generally can't just hop-on/hop-off without > being a menace to other traffic and there's sometimes even a physical > barrier; the way can also diverge from the main road way, taking > short-cuts round roundabouts or similar. And shortcuts still get their own highways since that's a point where they diverge from the main road. But for 99% of the cases where the cycleway just follows the road that's overkill. Nobody is going to have doubts about where the cycleway is exactly: next to the road. > > But there are more reasons why I don't like these as separate > > highways: > > > > * We're also not tagging sidewalks as separate "highway=footway" > > right (well, I guess there is not tag for sidewalks yet but it'll > > come -- but I can't imagine someone tagging them all like separate > > ways anyway, just think about the intersection mentioned above and > > add four "highway=footway"s to them). Cycleways are usually between > > the sidewalk and the road, so it becomes quite odd that a sidewalk > > is just a tag, but a cycleway is its own highway. > > That's just an argument for modeling the pavement properly. Most > pavements are just tacked onto the road as an extra "lane", but with > a kerb (to discourage the cars from using it :-) ), so I wouldn't > usually bother adding these as separate ways, but where the pavement > diverges or is clearly separate, it should probably be modelled as > such. But where is the point between being separated from the road and not? As a cyclist I can usually get off my bike and walk on the pavement next to it at any point. I can also cross the road at any point. A line of trees doesn't stop that. Some grass doesn't stop that, a line of parked cars doesn't either. Of course, at the exact location of a tree or perhaps a flower bed I can't hop from cycle lane to whatever, but I can't see a pavement being tagged like lots of small highways next to the road at every tree: /| \| | /| \| Anyway, the exact distinction between cycle lane and cycle track seems quite odd to me: in legislation there's no difference between them anyway, both give exactly the same rights/obligations as a road user (at least here in Belgium). So I guess everyone agrees that a cycleway which is painted right next to the part of the road where cars drive is a cycle lane. But at what point does it become a track? Does a line of 10cm high and 50cm long concrete blocks make it a cycle track? A 5m wide area painted diagonal stripes where no-one can drive between cycleway and "motorcarway"? A kerb? Parking spaces? A 50cm wide patch of grass? A line of trees? A 5m wide patch of grass? A flower bed? So where's the difference? I tend to see all of them as cycleway=track, except the painted diagonal stripes one. Yet I'd only consider making separate highway=cycleway ways when the distance becomes something like 10 meters. > > * It's just a lot harder to make them their own highways. it's > > much easier to make mistakes. > > It's possible to argue that one both ways: it's easier to see what's > going on and where cycle tracks start and stop, and where exactly > they are. At the zoom levels where you can easily see the separate cycle tracks (zoom level 18 or so) you could as well just have the renderer interpret the cycleway=* tags to lanes next to the road with arrows on them for exampl
Re: [OSM-talk] Cycle lanes
On Monday 24 March 2008 15:18:00 Dave Stubbs wrote: > There are always edge cases that cause problems. None of the things I mentioned above are really rare here in the Netherlands. I realise this country is not very big, so maybe we have to live with the fact we are an "edge country"? -- m.v.g., Cartinus ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
> > > * Rendering engines could handle it much easier if it were just a > > cycleway=* tag added to the road. > > Please show me the simple rendering algorithm for mapnik and osmarender you > have envisioned to make this working for all the special cases above. Until > you do, I keep believing the opposite is true, since it is fairly obvious > from the existence of Andy's cyclemap and the experimental Dutch cyclemap > that they have absolutely no problem rendering cycleways that are drawn as > separate ways. In general if Mapnik supported rendering a displaced line then a large number of the cases get covered. It already supports rendering line labels displaced, so it's not completely impossible. There are always edge cases that cause problems. And people reversing the direction of a way is a mistake, pure and simple... we have to consider way direction important otherwise oneway streets will never work... so hopefully people aren't going round randomly changing directions. For me the biggest current problem for rendering cycleways is indicating when a cyclist loses priority. Here in London there are a load of cycle tracks and shared paths where you lose priority over traffic from side streets if you're using it... this is really irritating so it's often better to cycle on the road. But on-road cyclelanes nearly always keep the priority as they are considered part of the main road, so it's worth knowing where they are. So we do need to fix this rendering issue anyway. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Mon, Mar 24, 2008 at 1:01 PM, Ben Laenen <[EMAIL PROTECTED]> wrote: > On Monday 24 March 2008, Andy Allan wrote: > > I think a lot of the physical cycleway tagging is ambiguous at the > > moment, especially with the cycleway= tag. I think cycleway=track was > > intended only for adding to highway=* (not highway=cycleway), but I > > would advise that all off-road cycle paths, including those on > > sidewalks, are drawn as a separate way with highway=cycleway instead. > > I beg to differ here. When you have to tag cycleways belonging to a road > not as "highway=whatever, cycleway=track" but as > separate "highway=cycleway" they just become an editing mess, > especially at intersections. Especially when adding route relations to > them. Just imagine two dual carriage ways with on either side a > cycleway crossing: you then need 24 different ways to just represent > that one intersection, like this: > > > ---- cycleway > ---- road > ---- road > ---- cycleway > I count 8 ways? Unless you are splitting all the ways at absolutely every intersection which is probably a little excessive. Obviously creating a way for every single cycle lane is going to just cause a mess, so where they do just follow the road, on the road, it's probably best to keep them as just a simple tag. However, where they are clearly separate, it's probably best to tag them as you would a dual-carriageway, and for the same reasons. With a separate cycleway you generally can't just hop-on/hop-off without being a menace to other traffic and there's sometimes even a physical barrier; the way can also diverge from the main road way, taking short-cuts round roundabouts or similar. > > But there are more reasons why I don't like these as separate highways: > > * We're also not tagging sidewalks as separate "highway=footway" right > (well, I guess there is not tag for sidewalks yet but it'll come -- but > I can't imagine someone tagging them all like separate ways anyway, > just think about the intersection mentioned above and add > four "highway=footway"s to them). Cycleways are usually between the > sidewalk and the road, so it becomes quite odd that a sidewalk is just > a tag, but a cycleway is its own highway. That's just an argument for modeling the pavement properly. Most pavements are just tacked onto the road as an extra "lane", but with a kerb (to discourage the cars from using it :-) ), so I wouldn't usually bother adding these as separate ways, but where the pavement diverges or is clearly separate, it should probably be modelled as such. > [snip] > > * It's just a lot harder to make them their own highways. it's much > easier to make mistakes. It's possible to argue that one both ways: it's easier to see what's going on and where cycle tracks start and stop, and where exactly they are. > > * Rendering engines could handle it much easier if it were just a > cycleway=* tag added to the road. >From practical experience I disagree. > > * You can usually arbitrarily go from the cycleway to the main road (to > cross it for example). Routing applications could make use of that, if > it's just a cycleway=* tag. Maybe you have to watch out for parked cars > for example, but I've seen cycle lanes where there are parked cars > between you and the road as well, yet the cycle lane is a lane and not > a track. (and before someone mentiones it: yes, relations like the > dual_carriage relation could solve that, but let us first get relation > support in editors a bit better before trying to put more and more into > relations) Potlatch is getting relation support sometime soon just awaiting deployment :-) Dave ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Monday 24 March 2008 14:01:59 Ben Laenen wrote: > On Monday 24 March 2008, Andy Allan wrote: > > I think a lot of the physical cycleway tagging is ambiguous at the > > moment, especially with the cycleway= tag. I think cycleway=track was > > intended only for adding to highway=* (not highway=cycleway), but I > > would advise that all off-road cycle paths, including those on > > sidewalks, are drawn as a separate way with highway=cycleway instead. > > I beg to differ here. When you have to tag cycleways belonging to a road > not as "highway=whatever, cycleway=track" but as > separate "highway=cycleway" they just become an editing mess, > especially at intersections. Especially when adding route relations to > them. Just imagine two dual carriage ways with on either side a > cycleway crossing: you then need 24 different ways to just represent > that one intersection, like this: > > > ---- cycleway > ---- road > ---- road > ---- cycleway > > > But there are more reasons why I don't like these as separate highways: > > * We're also not tagging sidewalks as separate "highway=footway" right > (well, I guess there is not tag for sidewalks yet but it'll come -- but > I can't imagine someone tagging them all like separate ways anyway, > just think about the intersection mentioned above and add > four "highway=footway"s to them). Cycleways are usually between the > sidewalk and the road, so it becomes quite odd that a sidewalk is just > a tag, but a cycleway is its own highway. > > * Here's the big argument: There's no information lost by adding tags to > the road like "cycleway=track" (we need a few more tags though for the > more exotic cases, like when the cycle track also serves as sidewalk), > compared to adding separate ways. So I'd like to keep the simplest way > then. And you need tags for when the cycleway on the left is bi-directional, but on the right it's one-way. Or tags for when there is a cycleway only on one side. I suppose you want this to be done with tags/values like left_track or left:track etc. Which becomes a mess when somebody reverses the direction of the way without changing all the tags. Not to mention special turn restrictions for bicycles only, if you can't turn left with your bicycle, but you can with your car. etc., etc., etc., Yes real easy. > * It's just a lot harder to make them their own highways. it's much > easier to make mistakes. Since it is impossible to make all the special cases above visible at a glance in an editor the opposite is true. > * Rendering engines could handle it much easier if it were just a > cycleway=* tag added to the road. Please show me the simple rendering algorithm for mapnik and osmarender you have envisioned to make this working for all the special cases above. Until you do, I keep believing the opposite is true, since it is fairly obvious from the existence of Andy's cyclemap and the experimental Dutch cyclemap that they have absolutely no problem rendering cycleways that are drawn as separate ways. > * You can usually arbitrarily go from the cycleway to the main road (to > cross it for example). Where I live there are numerous places where this is not the case. > Routing applications could make use of that, if > it's just a cycleway=* tag. Maybe you have to watch out for parked cars > for example, but I've seen cycle lanes where there are parked cars > between you and the road as well, yet the cycle lane is a lane and not > a track. (and before someone mentiones it: yes, relations like the > dual_carriage relation could solve that, but let us first get relation > support in editors a bit better before trying to put more and more into > relations) > > * Routing hell, like mentioned above: adding cycle routes to an > intersection with highway=cycleway next to highway=whatever will make a > simple route crossing the intersection a patch of a dozen ways > belonging to the route. Also for simple roads which aren't dual > carriage, you're then forced to add two ways (one in each direction) > instead of just the main road to the route relation. > > I've tagged the cycleways as their own highway once, but just doing that > I got quickly convinced that doing that was just a bad idea, so I > stopped and even just reverted everything to cycleway=* tags. So far I > haven't heard one argument which invalidates the issues above enough > that it convinced me highway=cycleway for these is a good idea. -- m.v.g., Cartinus ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Monday 24 March 2008, Andy Allan wrote: > I think a lot of the physical cycleway tagging is ambiguous at the > moment, especially with the cycleway= tag. I think cycleway=track was > intended only for adding to highway=* (not highway=cycleway), but I > would advise that all off-road cycle paths, including those on > sidewalks, are drawn as a separate way with highway=cycleway instead. I beg to differ here. When you have to tag cycleways belonging to a road not as "highway=whatever, cycleway=track" but as separate "highway=cycleway" they just become an editing mess, especially at intersections. Especially when adding route relations to them. Just imagine two dual carriage ways with on either side a cycleway crossing: you then need 24 different ways to just represent that one intersection, like this: ---- cycleway ---- road ---- road ---- cycleway But there are more reasons why I don't like these as separate highways: * We're also not tagging sidewalks as separate "highway=footway" right (well, I guess there is not tag for sidewalks yet but it'll come -- but I can't imagine someone tagging them all like separate ways anyway, just think about the intersection mentioned above and add four "highway=footway"s to them). Cycleways are usually between the sidewalk and the road, so it becomes quite odd that a sidewalk is just a tag, but a cycleway is its own highway. * Here's the big argument: There's no information lost by adding tags to the road like "cycleway=track" (we need a few more tags though for the more exotic cases, like when the cycle track also serves as sidewalk), compared to adding separate ways. So I'd like to keep the simplest way then. * It's just a lot harder to make them their own highways. it's much easier to make mistakes. * Rendering engines could handle it much easier if it were just a cycleway=* tag added to the road. * You can usually arbitrarily go from the cycleway to the main road (to cross it for example). Routing applications could make use of that, if it's just a cycleway=* tag. Maybe you have to watch out for parked cars for example, but I've seen cycle lanes where there are parked cars between you and the road as well, yet the cycle lane is a lane and not a track. (and before someone mentiones it: yes, relations like the dual_carriage relation could solve that, but let us first get relation support in editors a bit better before trying to put more and more into relations) * Routing hell, like mentioned above: adding cycle routes to an intersection with highway=cycleway next to highway=whatever will make a simple route crossing the intersection a patch of a dozen ways belonging to the route. Also for simple roads which aren't dual carriage, you're then forced to add two ways (one in each direction) instead of just the main road to the route relation. I've tagged the cycleways as their own highway once, but just doing that I got quickly convinced that doing that was just a bad idea, so I stopped and even just reverted everything to cycleway=* tags. So far I haven't heard one argument which invalidates the issues above enough that it convinced me highway=cycleway for these is a good idea. Greetings Ben ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
Sounds very similar to the cycleway tagging in Bedford; treat it as a separate way if it's not on the road, which makes it easy to show if it takes detours away from the road: http://informationfreeway.org/?lat=52.1218000936748&lon=-0.489018411255609&zoom=17&layers=F0B0F (the gaps are where it merges with a footpath; it's assumed that these exist at the sides of most roads, although it might be useful to tag sections where you're specifically permitted to cycle on the path but have to give priority to pedestrians) On Mon, Mar 24, 2008 at 7:48 AM, Mike Collinson <[EMAIL PROTECTED]> wrote: > At 07:32 AM 24/03/2008, Lars Aronsson wrote: > > >It's common in Sweden to have wide sidewalks divided into half > >footway, half cycleway. This can happen on either or both sides > >of the street. Should this be tagged as highway=*;cycleway=lane? > >Technically speaking it isn't a lane because it's above the curb. > > > >How can I indicate which (or both) sides of the street it applies > >to? If it's only on one side, how can I indicate this? > > > >The page [[Proposed features/Sidewalk]] proposes sidewalk=right. > >Should this be used for cycleway too? A user comment on that page > >mentions cycleway:right=lane. Is that a common use, that needs to > >be documented? > > > >The page [[Cycleway]] instead talks of lane_left and lane_right. > > > >The same page also mentions the "width" key, but what are some > >useful values for this key? In the case of two wide sidewalks > >that allow bikes, some separating grass, and double street lanes, > >should the width be the total width (in metres) from wall to wall? > > As a cyclist currently in Stockholm, I personally draw a separate way and > label it highway=cycleway, cycleway=track, foot=yes, surface=paved. > > The main rationale for that is that it is a separate track rather than a > painted lane division on someone else's road. Sometimes the track is right > next to the road, sometimes separated by a grass verge and sometimes it > wanders off into the woods. > > But my main reason for doing that is safety. I *much* prefer riding these > cycleways than on lanes marked on the road and deliberately plan my routes > according. Drawing a separate way makes them very obvious on the map using > existing tagging. I can also easily see where they change side of a road > and if they have sections where they merge into the normal road as a lane or > just disappear. > > That is my personal opinion, I'm interested it what others think. > > Mike > > PS Here is a work-in-progress example: > > Good: The southbound Sankt Eriksgatan sidewalk cycleway really is a > separate track over the bridge and really does abruptly end at > Aströmergatan. As a bonus, it is clearly shown as one-way. > > Bad: An east-west sidewalk cycleway begins at the corner of Fridhemsgatan > and Drottingholmvägen, but is difficult to see because of the rendering. > > > http://www.openstreetmap.org/?lat=59.33395&lon=18.02934&zoom=17&layers=B0FT > > > > ___ > talk mailing list > talk@openstreetmap.org > http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk > ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
On Monday 24 March 2008 09:53:07 Andy Allan wrote: > If the way is tagged with highway=cycleway I don't think it needs > cycleway=track, btw. +1 > I'm going to put together a guide for how to tag cycle paths, since > I've been contacted by a few other groups who are finding our tagging > insufficient for their needs (such as shared use vs segregated paths), > and I hope that'll clear things up a bit. Great > On Mon, Mar 24, 2008 at 7:48 AM, Mike Collinson <[EMAIL PROTECTED]> wrote: > > As a cyclist currently in Stockholm, I personally draw a separate way > > and label it highway=cycleway, cycleway=track, foot=yes, surface=paved. I think the surface=paved is redundant too, because that is what I would expect if no surface tag was put on the way. -- just my 2c, Cartinus ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
I think a lot of the physical cycleway tagging is ambiguous at the moment, especially with the cycleway= tag. I think cycleway=track was intended only for adding to highway=* (not highway=cycleway), but I would advise that all off-road cycle paths, including those on sidewalks, are drawn as a separate way with highway=cycleway instead. If the way is tagged with highway=cycleway I don't think it needs cycleway=track, btw. I'm going to put together a guide for how to tag cycle paths, since I've been contacted by a few other groups who are finding our tagging insufficient for their needs (such as shared use vs segregated paths), and I hope that'll clear things up a bit. Certainly the easiest for now is that if it's not on the road, it's a highway=cycleway. Cheers, Andy On Mon, Mar 24, 2008 at 7:48 AM, Mike Collinson <[EMAIL PROTECTED]> wrote: > At 07:32 AM 24/03/2008, Lars Aronsson wrote: > > >It's common in Sweden to have wide sidewalks divided into half > >footway, half cycleway. This can happen on either or both sides > >of the street. Should this be tagged as highway=*;cycleway=lane? > >Technically speaking it isn't a lane because it's above the curb. > > > >How can I indicate which (or both) sides of the street it applies > >to? If it's only on one side, how can I indicate this? > > > >The page [[Proposed features/Sidewalk]] proposes sidewalk=right. > >Should this be used for cycleway too? A user comment on that page > >mentions cycleway:right=lane. Is that a common use, that needs to > >be documented? > > > >The page [[Cycleway]] instead talks of lane_left and lane_right. > > > >The same page also mentions the "width" key, but what are some > >useful values for this key? In the case of two wide sidewalks > >that allow bikes, some separating grass, and double street lanes, > >should the width be the total width (in metres) from wall to wall? > > As a cyclist currently in Stockholm, I personally draw a separate way and > label it highway=cycleway, cycleway=track, foot=yes, surface=paved. > > The main rationale for that is that it is a separate track rather than a > painted lane division on someone else's road. Sometimes the track is right > next to the road, sometimes separated by a grass verge and sometimes it > wanders off into the woods. > > But my main reason for doing that is safety. I *much* prefer riding these > cycleways than on lanes marked on the road and deliberately plan my routes > according. Drawing a separate way makes them very obvious on the map using > existing tagging. I can also easily see where they change side of a road and > if they have sections where they merge into the normal road as a lane or just > disappear. > > That is my personal opinion, I'm interested it what others think. > > Mike > > PS Here is a work-in-progress example: > > Good: The southbound Sankt Eriksgatan sidewalk cycleway really is a > separate track over the bridge and really does abruptly end at Aströmergatan. > As a bonus, it is clearly shown as one-way. > > Bad: An east-west sidewalk cycleway begins at the corner of Fridhemsgatan > and Drottingholmvägen, but is difficult to see because of the rendering. > > http://www.openstreetmap.org/?lat=59.33395&lon=18.02934&zoom=17&layers=B0FT > > > > > > ___ > talk mailing list > talk@openstreetmap.org > http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk > ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
At 07:32 AM 24/03/2008, Lars Aronsson wrote: >It's common in Sweden to have wide sidewalks divided into half >footway, half cycleway. This can happen on either or both sides >of the street. Should this be tagged as highway=*;cycleway=lane? >Technically speaking it isn't a lane because it's above the curb. > >How can I indicate which (or both) sides of the street it applies >to? If it's only on one side, how can I indicate this? > >The page [[Proposed features/Sidewalk]] proposes sidewalk=right. >Should this be used for cycleway too? A user comment on that page >mentions cycleway:right=lane. Is that a common use, that needs to >be documented? > >The page [[Cycleway]] instead talks of lane_left and lane_right. > >The same page also mentions the "width" key, but what are some >useful values for this key? In the case of two wide sidewalks >that allow bikes, some separating grass, and double street lanes, >should the width be the total width (in metres) from wall to wall? As a cyclist currently in Stockholm, I personally draw a separate way and label it highway=cycleway, cycleway=track, foot=yes, surface=paved. The main rationale for that is that it is a separate track rather than a painted lane division on someone else's road. Sometimes the track is right next to the road, sometimes separated by a grass verge and sometimes it wanders off into the woods. But my main reason for doing that is safety. I *much* prefer riding these cycleways than on lanes marked on the road and deliberately plan my routes according. Drawing a separate way makes them very obvious on the map using existing tagging. I can also easily see where they change side of a road and if they have sections where they merge into the normal road as a lane or just disappear. That is my personal opinion, I'm interested it what others think. Mike PS Here is a work-in-progress example: Good: The southbound Sankt Eriksgatan sidewalk cycleway really is a separate track over the bridge and really does abruptly end at Aströmergatan. As a bonus, it is clearly shown as one-way. Bad: An east-west sidewalk cycleway begins at the corner of Fridhemsgatan and Drottingholmvägen, but is difficult to see because of the rendering. http://www.openstreetmap.org/?lat=59.33395&lon=18.02934&zoom=17&layers=B0FT ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Cycle lanes
> It's common in Sweden to have wide sidewalks divided into half > footway, half cycleway. This can happen on either or both sides > of the street. Should this be tagged as highway=*;cycleway=lane? > Technically speaking it isn't a lane because it's above the curb. Theoretically, you could use "track", but in the long run, that's not enough. > How can I indicate which (or both) sides of the street it applies > to? If it's only on one side, how can I indicate this? Map the "sidewalks" separately from the road as "highway=cycleway foot=yes" etc. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk