Re: [OSM-talk] Divided roads proposal
On Mon, Dec 7, 2009 at 9:12 AM, Richard Bullock rb...@cantab.net wrote: Maybe we should be mapping slipways, hopefully there's a better approach than marking them all as fully fledged roads though. Sliproads are tagged as highway=xyz_link e.g. a sliproad to a motorway would be highway=motorway_link sliproad to a trunk road would be highway=trunk_link etc. Ah, I didn't know it went all the way down to secondary_link. (But not tertiary_link?) Can renderers improve their render quality at lower zoom levels by not rendering (certain) link roads? Ie, given road A-B-C, with incoming road D-B, and link D-A, perhaps it could not render D-A. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Mon, Dec 7, 2009 at 7:34 AM, Steve Bennett stevag...@gmail.com wrote: Can renderers improve their render quality at lower zoom levels by not rendering (certain) link roads? Ie, given road A-B-C, with incoming road D-B, and link D-A, perhaps it could not render D-A. Cartographers (people) can do most anything. Renderers (software) are pretty good too. ;-) In Mapnik, for example, you could select by zoom-level to either render or not render all highway=secondary_link. This is a rendering rule from Mapnik for secondary_link (secondary as well) when zoomed all the way in (from 1:1000 to 1:5000). It shows that the secondary_link will be rendered as 17 pixels wide. Zooming out causes the width to reduce to 12 pixels, 10 pixels, 4 pixels then disappear at scales beyond 1:15 Rule Filter[highway] = 'secondary' or [highway] = 'secondary_link'/Filter MaxScaleDenominator5000/MaxScaleDenominator MinScaleDenominator1000/MinScaleDenominator LineSymbolizer CssParameter name=stroke#a37b48/CssParameter CssParameter name=stroke-width17/CssParameter CssParameter name=stroke-dasharray4,2/CssParameter /LineSymbolizer /Rule You could easily choose to not show secondary_link at scales of your choice. Whether that is an improvement in rendering quality or not would be a judgment call and should consider the intent of your rendering and the interests of your audience. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Tue, Dec 8, 2009 at 12:06 AM, Richard Weait rich...@weait.com wrote: You could easily choose to not show secondary_link at scales of your choice. Whether that is an improvement in rendering quality or not would be a judgment call and should consider the intent of your rendering and the interests of your audience. Ok, but I was really thinking about the standard Mapnik run at OSM. And I think it would take more than simply not rendering *any* trunk_links, for example. It would have to check that the link was actually redundant: I think links that do a 360 or cross another road should always be shown, and probably all motorway_links for that matter. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Sat, Dec 5, 2009 at 1:43 AM, Anthony o...@inbox.org wrote: 1, 2. Dual carriageway 5, 6, 7, 8, 9, 10. Dual carriageway Alright, but let's be practical. It's a lot of effort to create and maintain pairs of roads (let's not call them dual carriageways - that's really a specific type of motorway). Let's imagine this tag is implemented and there is renderer support. What value do you see in mapping examples 2 and 10 as pairs of roads rather than a single road with divided=median? Is the benefit just so you can get more precise with area micromapping? Let's assume, because it's true, that volunteer mapping time is limited, and the use of areas to micromap roads is rare, and certainly not expected by end users. Why is a pair of roads better in 2 than a single, divided road? Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Steve Bennett wrote: On Sat, Dec 5, 2009 at 1:43 AM, Anthony o...@inbox.org mailto:o...@inbox.org wrote: 1, 2. Dual carriageway 5, 6, 7, 8, 9, 10. Dual carriageway Alright, but let's be practical. It's a lot of effort to create and maintain pairs of roads (let's not call them dual carriageways - that's really a specific type of motorway). Let's imagine this tag is implemented and there is renderer support. What value do you see in mapping examples 2 and 10 as pairs of roads rather than a single road with divided=median? Is the benefit just so you can get more precise with area micromapping? Let's assume, because it's true, that volunteer mapping time is limited, and the use of areas to micromap roads is rare, and certainly not expected by end users. Why is a pair of roads better in 2 than a single, divided road? Many of the example pictures have end cases that need to be handled by separated roads, so why not just draw the reality on the ground? Having to add MORE complexity such as you can turn off this side of the road is just as wrong? So map the dual carriage way section, and show the other detail joining to the physical situation? In many areas, the macro level view is complete and people are starting to look at the micro view. Just because a few areas 'might look better' in the macro view is no reason to remove the existing format, and I see little point adding tags for something that is hiding the micro view? SOME roads do need a 'divider' tag, but only to add 'white line' and other 'micro' data that can't be included by areas or other means. The crosshatch area is just another edge case that needs to be handled in bother levels. -- 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] Divided roads proposal
On Sun, Dec 6, 2009 at 10:43 PM, Lester Caine les...@lsces.co.uk wrote: Many of the example pictures have end cases that need to be handled by separated roads, so why not just draw the reality on the ground? Because, as someone else pointed out, drawing the reality on the ground isn't the only, or even, best approach for mapping. The slightly more abstract this is a single road, divided works better than this is two roads in some cases. It seems to me that to almost any proposal you could argue why not just draw the reality on the ground?: Q: Can we have a cinema tag? A: Why not just draw the reality on the ground? It's a building, with chairs, a room with a projector... Q: Can we have a swimming pool tag? A: Why not just draw the reality on the ground? It's a waterway, surface=tiles, foot=yes, bicycle=no... Having to add MORE complexity such as you can turn off this side of the road is just as wrong? I don't think a single road with a single junction leading to a side road, with a single tag, could possibly be construed as more complex than two roads with an extra road for the gap. In many areas, the macro level view is complete and people are starting to look at the micro view. Just because a few areas 'might look better' in the macro view is no reason to remove the existing format, I repeat: I am not proposing removing anything. I'm proposing *adding* a tag, primarily to allow *adding* information about *new* streets. It's possible this will mean that some streets that *would have* been mapped as two roads will instead be mapped as one road, but it's a stretch to call that removing anything. (Sorry for the irritated tone, but...c'mon.) and I see little point adding tags for something that is hiding the micro view? If and when roads are mapped as areas (in addition to ways), I'm sure you'll be able to map the individual halves of the divided road to your heart's content. Just like you'll be able to map every lane, every slipway, every traffic island and every painted arrow. Honestly though, the primary function of OSM is not the micro view. We're not primarily interested in centimetre perfect placement of lumps of concrete. Are you really suggesting that we don't use a feature because it might interfere with the micro view, even though it works better at the macro view? Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Steve Bennett wrote: Honestly though, the primary function of OSM is not the micro view. We're not primarily interested in centimetre perfect placement of lumps of concrete. Are you really suggesting that we don't use a feature because it might interfere with the micro view, even though it works better at the macro view? Why do you say THAT ? Many of the additions currently being discussed ARE because the macro view is now complete, and adding the fine detail is now being carried out. There is no PRIMARY interest. Everybody has their own views on what is important! I think my only problem with 'divided' is At what point do you apply it? The samples being shown are quite clearly - on the whole - dual carriageway structures. Example 10 clearly has a more complex structure than can be mapped by showing a 'divided' tag, since there is no access to the joining road from the other carriageway? Example 6 has some quite complex slip roads that really need isolated ways for the main carriageway. Trying to ADD tags to supplement a simple 'divided' tag to explain the slips on and off at the end is handled much easier with a simple dual carriage way? And many of the other examples need the same end cases. So at what level does a simple 'divided' tag actually work in practice? However 'double white lines' on a single carriage way road IS a divider that needs tagging? -- 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] Divided roads proposal
On Mon, Dec 7, 2009 at 12:30 AM, Lester Caine les...@lsces.co.uk wrote: I think my only problem with 'divided' is At what point do you apply it? The samples being shown are quite clearly - on the whole - dual carriageway structures. (Just on terminology, I'm used to dual carriageway only being applied to motorways, but Wikipedia says it technically applies to any road. We'll go with that, then, ok.) By are clearly dual carriageway structures, I take it you're distinguishing between roads which have a shortish traffic island of some type, versus those which are divided for a long stretch. Is this important? Example 10 clearly has a more complex structure than can be mapped by showing a 'divided' tag, since there is no access to the joining road from the other carriageway? Have you read the proposal? http://wiki.openstreetmap.org/wiki/Proposed_features/Divided_road It covers *exactly* this case. In my proposal (ie, not Vovanium's, which cohabits the page), when there's a junction (like the road entering from SW), by default that road can't cross the median - there is no gap. You can see how this looks in my mocked up Halcyon image. Example 6 has some quite complex slip roads that really need isolated ways for the main carriageway. Trying to ADD tags to supplement a simple 'divided' tag to explain the slips on and off at the end is handled much easier with a simple dual carriage way? Do slip ways need to be modelled at all? At the moment, they're not, as far as I have seen. Essentially what is going on at that intersection is very straight forward: divided road meets (temporarily) divided road, and all turns are possible. Currently, I don't think many people would map the N/S as divided (ie, two ways). With this proposal, you could do so, without creating a mess. And many of the other examples need the same end cases. So at what level does a simple 'divided' tag actually work in practice? However IMHO, the divided tag is well suited to all cases except 6, 7, and 9. Number 2 in particular is a perfect example. Worth tagging, not worth splitting the road in two for. 'double white lines' on a single carriage way road IS a divider that needs tagging? Yeah, I deliberated over whether or not to include that. What do you think - same proposal, or separate? Double white lines really aren't a divider, they're a restriction on overtaking. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Sun, Dec 6, 2009 at 5:55 AM, Steve Bennett stevag...@gmail.com wrote: On Sat, Dec 5, 2009 at 1:43 AM, Anthony o...@inbox.org wrote: 1, 2. Dual carriageway 5, 6, 7, 8, 9, 10. Dual carriageway Alright, but let's be practical. It's a lot of effort to create and maintain pairs of roads (let's not call them dual carriageways - that's really a specific type of motorway). Let's imagine this tag is implemented and there is renderer support. What value do you see in mapping examples 2 and 10 as pairs of roads rather than a single road with divided=median? Is the benefit just so you can get more precise with area micromapping? Let's assume, because it's true, that volunteer mapping time is limited, and the use of areas to micromap roads is rare, and certainly not expected by end users. Why is a pair of roads better in 2 than a single, divided road? If you don't see how it's more accurate, I can't help you. Yes, in many of those cases you outline, it's overkill. But in those same cases, just leaving a single way and not worrying about the divider at all is fine. Only in a case where the divider provides routing information (other than the ability to U-turn) would I say that it's important to use a dual carriageway (seems to me to fit the dictionary definition). The divided=median tag is already micromapping. All I'm saying is if you're going to start micromapping, do it right. OTOH, if you don't intend to use divided=* to represent routing information, so routers and renderers can safely ignore the tag, then do whatever you want with it. Like I suggested, I'll just treat it as a type of todo tag. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Mon, Dec 7, 2009 at 1:03 AM, Anthony o...@inbox.org wrote: Is the benefit just so you can get more precise with area micromapping? Let's assume, because it's true, that volunteer mapping time is limited, and the use of areas to micromap roads is rare, and certainly not expected by end users. Why is a pair of roads better in 2 than a single, divided road? If you don't see how it's more accurate, I can't help you. I said better. You say accurate, but you probably mean precise. Let's say I ask you the time. You reply Sun Dec 6 06:11:07 2009 PST. A better answer would have been ten past one AM. I can understand your temptation to think that increasing levels of precision are always better: usually it is. But the problem here is: 1) The extra precision of precisely mapping two ways means extra complexity, hence more difficulty in rendering, and it maps less well onto the user's mental model. Just like amenity=swimming_pool is better, if less precise, than a natural=water, with precise details of surface materials and whatnot. 2) Extra precision requires more information, which we don't necessarily have. How would you map a divided road which you don't have an aerial photo for? You have a GPS trace with points every 10 metres, with an accuracy of about 5 metres. How would you map this: two ways? Now you see the difference between precision and accuracy: you have precisely mapped out two ways, when in fact neither is particularly accurate. 3) Extra precision requires more time, which we don't necessarily have. Let's say that we agree that all divided roads should be mapped as two ways. Let's also say it takes on average 1 minute to trace out a single way. There's a volunteer with 60 divided roads to map in front of him, and he's got one hour to spend. See where I'm going? Yes, in many of those cases you outline, it's overkill. But in those same cases, just leaving a single way and not worrying about the divider at all is fine. Absolutely. We agree on two situations: 1) Roads with a division too trivial to map, which we map as a single way. 3) Large, dual-carriage roads with a division too complex to model with a simple divided=* tag, which we map as two separate ways. Can you guess what number 2) is? Only in a case where the divider provides routing information (other than the ability to U-turn) would I say that it's important to use a dual carriageway (seems to me to fit the dictionary definition). The divided=median tag is already micromapping. I disagree, but I accept that the line of what is considered to be micromapping is subjective. You explain very well the current conundrum: there are currently only two choices, either map out the division as two separate roads (a dual carriageway), or declare it micromapping and ignore it. I'm proposing a third choice, that lets you capture some useful information without the overhead of the dual carriageway. All I'm saying is if you're going to start micromapping, do it right. This is the most compelling argument against my proposal, and I'm surprised no one has brought it up yet: the divider=* tag can be used for certain kinds of traffic islands (namely those that run down the middle of a two way road), but not others (such as slipway dividers, islands in one way streets, islands in intersections...) However, I think do it right is an immense, open-ended, complex task. This proposal addreses enough common situations that it's worth implementing, even if we still can't map *everything*. OTOH, if you don't intend to use divided=* to represent routing information, so routers and renderers can safely ignore the tag, then do whatever you want with it. Like I suggested, I'll just treat it as a type of todo tag. I suggest you read the Routing section of the proposal. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Steve Bennett wrote: On Mon, Dec 7, 2009 at 12:30 AM, Lester Caine les...@lsces.co.uk mailto:les...@lsces.co.uk wrote: I think my only problem with 'divided' is At what point do you apply it? The samples being shown are quite clearly - on the whole - dual carriageway structures. (Just on terminology, I'm used to dual carriageway only being applied to motorways, but Wikipedia says it technically applies to any road. We'll go with that, then, ok.) MANY major routes in the UK are trunk roads and most routes around cities will have one way elements that split and join at different points. YES it is the A?? and a single name, but the structure can only be mapped as a dual carriage way. Which then takes us to some of the 'green way' areas where cars go down either side of a grass verge. A simple 'divided' in your tag, but the separation may grow from nothing to several meters. At what point do you change from 'divided' to separate ways, which then begs the question - why have divided if it's just a shorthand for two ways with opposite directions. By are clearly dual carriageway structures, I take it you're distinguishing between roads which have a shortish traffic island of some type, versus those which are divided for a long stretch. Is this important? Example 10 clearly has a more complex structure than can be mapped by showing a 'divided' tag, since there is no access to the joining road from the other carriageway? Have you read the proposal? http://wiki.openstreetmap.org/wiki/Proposed_features/Divided_road You need to justify the real need for it. I'll continue to map the actual structure, and add the additional ways for the related footpaths. I don't see the need for this shorthand for many of the cases you are trying to make? It covers *exactly* this case. In my proposal (ie, not Vovanium's, which cohabits the page), when there's a junction (like the road entering from SW), by default that road can't cross the median - there is no gap. You can see how this looks in my mocked up Halcyon image. I think what YOU are missing is that in most cases where there are traffic islands which add one way sections of way, they ARE mapped. Around here there was an attempted to remove some of them, but that has been rolled back, so where a road splits, the correct direction ways are added. Routing does not then need to run through lots of additional tags to find if it can then do a maneuver ... Example 6 has some quite complex slip roads that really need isolated ways for the main carriageway. Trying to ADD tags to supplement a simple 'divided' tag to explain the slips on and off at the end is handled much easier with a simple dual carriage way? Do slip ways need to be modelled at all? At the moment, they're not, as far as I have seen. Essentially what is going on at that intersection is very straight forward: divided road meets (temporarily) divided road, and all turns are possible. Currently, I don't think many people would map the N/S as divided (ie, two ways). With this proposal, you could do so, without creating a mess. I think it is essential that slipways are mapped. ESPECIALLY when one is trying to add the right routing instructions. TomTom has started showing motorway and major road slipway details properly. You need to know when to get to an inside lane and take a slip road PRIOR to the actual junction. These are no different to the island details approaching a roundabout, so trying to 'save time' by not actually adding quite important detail does seem wrong? A tag saying you should have taken the slip 10 mts before this junction is not sensible. And many of the other examples need the same end cases. So at what level does a simple 'divided' tag actually work in practice? However IMHO, the divided tag is well suited to all cases except 6, 7, and 9. Number 2 in particular is a perfect example. Worth tagging, not worth splitting the road in two for. 'double white lines' on a single carriage way road IS a divider that needs tagging? Yeah, I deliberated over whether or not to include that. What do you think - same proposal, or separate? Double white lines really aren't a divider, they're a restriction on overtaking. Example 3 is no more than a wide 'double line' road marking. SO is it a 'divided' or is it simply a road marking? The problem with the proposal is that it does not have any indication on when it should be used ... or when the more detailed current methods are actually more 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
Re: [OSM-talk] Divided roads proposal
3) Extra precision requires more time, which we don't necessarily have. Let's say that we agree that all divided roads should be mapped as two ways. Let's also say it takes on average 1 minute to trace out a single way. There's a volunteer with 60 divided roads to map in front of him, and he's got one hour to spend. See where I'm going? And you don't have to do all 60 dual carriageways in one sitting. But to be completely honest, mapping out dual carriageways is really not *that* time consuming. In JOSM you could just copy the way you have drawn and drag the copied way a few metres to the side and reverse the direction. You'd then have to connect up side roads and tweak a few nodes - but should take no more than a couple of seconds if you can't be bothered drawing out the other carriageway. I'm sure Potlatch probably has a similar feature. And in a world with a large but finite number of roads, where a relatively small fraction of roads are dual carriageway, and an ever increasing army of mappers - it's really not that bad. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Mon, Dec 7, 2009 at 4:11 AM, Richard Bullock rb...@cantab.net wrote: But to be completely honest, mapping out dual carriageways is really not *that* time consuming. In JOSM you could just copy the way you have drawn and drag the copied way a few metres to the side and reverse the direction. If that's the case, it sort of shoots down the dual carriageway contains more information than a single way argument, doesn't it? You'd then have to connect up side roads and tweak a few nodes - but should take no more than a couple of seconds if you can't be bothered drawing out the other carriageway. I'm sure Potlatch probably has a similar feature. I'm not sure it does. And in a world with a large but finite number of roads, where a relatively small fraction of roads are dual carriageway, and an ever increasing army of mappers - it's really not that bad. Speaking for myself, what motivated my interest in this is that it is tedious, and doesn't feel like the right thing to be doing. These minor roads don't seem like they deserve too complete roads. Google Maps maps them as single roads. Melway maps them as a single road which renders like a double road (ie, variations in median width aren't accounted for). Yahoo — now that I look — does exactly what I'm proposing, rendering a single road with a dotted line down the middle to indicate the division. Now, I don't know what weight what everyone else is doing carries, but it indicates something. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Mon, Dec 7, 2009 at 2:31 AM, Lester Caine les...@lsces.co.uk wrote: grow from nothing to several meters. At what point do you change from 'divided' to separate ways, which then begs the question This is the same kind of question as when a road switches from tertiary to secondary etc. Does it need a firm answer? It would only be problematic, IMHO, if a road was frequently switching from one form to the other. - why have divided if it's just a shorthand for two ways with opposite directions. I don't think it is just a shorthand. In many cases I would put it the other way: two ways with opposite directions is just a kludge for a single divided road. Up until now, we haven't had a way to properly tag a single divided road, so we do it with two ways. We can fix that. You need to justify the real need for it. I'll continue to map the actual structure, and add the additional ways for the related footpaths. I don't see the need for this shorthand for many of the cases you are trying to make? If you're mapping every footpath, you are clearly working at a different level of detail. It's great that your area apparently has every road already covered already. Mine doesn't. To put it differently, the real need is to be able to efficiciently map areas which have divided roads, without resorting to micromapping them as two ways. I understand that you wouldn't use this tag. But would you object to others using it? I think what YOU are missing is that in most cases where there are traffic islands which add one way sections of way, they ARE mapped. Around here there Yes. And does mapping a traffic island as a splitting of a road into two ways not scream wrong! at you? Example: http://www.openstreetmap.org/?lat=-37.820693lon=144.919989zoom=18layers=B000FTF This looks right to you? was an attempted to remove some of them, but that has been rolled back, so where a road splits, the correct direction ways are added. Routing does not then need to run through lots of additional tags to find if it can then do a maneuver ... One key for ways, one key for junctions. Trivial stuff. I think it is essential that slipways are mapped. ESPECIALLY when one is trying to add the right routing instructions. TomTom has started showing motorway and major road slipway details properly. You need to know when to get to an inside lane and take a slip road PRIOR to the actual junction. Maybe we should be mapping slipways, hopefully there's a better approach than marking them all as fully fledged roads though. These are no different to the island details approaching a roundabout, so trying to 'save time' by not actually adding quite important detail does seem wrong? Even if it were the case that this was important detail which I was proposing not adding, it wouldn't be wrong, because surely a slipway here is not as important as a whole road not mapped somewhere else. Example 3 is no more than a wide 'double line' road marking. SO is it a 'divided' or is it simply a road marking? Either way, we currently have no way of marking it. The problem with the proposal is that it does not have any indication on when it should be used ... Other than, say, http://wiki.openstreetmap.org/wiki/Proposed_features/Divided_road#Scope Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
But to be completely honest, mapping out dual carriageways is really not *that* time consuming. In JOSM you could just copy the way you have drawn and drag the copied way a few metres to the side and reverse the direction. If that's the case, it sort of shoots down the dual carriageway contains more information than a single way argument, doesn't it? Speaking for myself, what motivated my interest in this is that it is tedious, and doesn't feel like the right thing to be doing. These minor roads don't seem like they deserve too complete roads No, it doesn't shoot down the argument. I just said it was an option if you couldn't be bothered to map it properly. And really, it doesn't take long anyway. And your proposed scheme for routing makes your scheme more complicated for the routers than mapping as two ways. In the existing scheme, if a shared node is not present between two ways, no routing is possible directly between the two. This is true regardless of what routing software you are using - it doesn't require extra tags or fudges to make it work. And think for a minute that there might be external data users who might not know to update their routing software for these extra tags. And not knowing about physically impossible turns is worse than not knowing about legal restrictions (turn restrictions). I can't disagree more about minor roads not deserving to be mapped fully. We should map everything to the best of our abilities. And given the community we have, I'm sure someone will sooner or later - even if you find it 'tedious'. Some people actually enjoy mapping you know... Google Maps maps them as single roads. Not around here. http://tinyurl.com/yz8y8dh The white coloured roads are just roads for a shopping centre. Scroll a bit further south, and you'll even see where G-Maps have split the road around a long-ish traffic island. A bit further south still, is a motorway junction where Google has mapped sliproads (and even marked on a sliproad bypassing the roundabout). The Yahoo for the same location http://uk.maps.yahoo.com/#mvt=mlat=53.465566lon=-2.354163zoom=18 The principle we have used in the past is where it is not physically possible to cross between two carriageways without leaving the road surface then mark it as two separate ways. I've not seen a compelling argument against continuing this practice. The only argument I've seen from you is essentially, it's boring - it takes far too long - which I don't really agree with. Dual carriageways all over the world are mapped like this on OSM (and it seems Google, Yahoo and I'm sure others do as well). ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Maybe we should be mapping slipways, hopefully there's a better approach than marking them all as fully fledged roads though. Sliproads are tagged as highway=xyz_link e.g. a sliproad to a motorway would be highway=motorway_link sliproad to a trunk road would be highway=trunk_link etc. See the wiki map-features page. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
( Forgot to FIX the address :( Another crap practice in OSM! ) Richard Bullock wrote: The principle we have used in the past is where it is not physically possible to cross between two carriageways without leaving the road surface then mark it as two separate ways. I've not seen a compelling argument against continuing this practice. The only argument I've seen from you is essentially, it's boring - it takes far too long - which I don't really agree with. Dual carriageways all over the world are mapped like this on OSM (and it seems Google, Yahoo and I'm sure others do as well). And that is what I would agree is the FIRST choice for mapping any divided structure. You do not bodge the detail so that it renders correctly, the rendering needs fixing if it is not displayed properly. Trying to add more tags for situations that ARE already covered by normal mapping practice is wrong. Adding a tag to cover details that ARE missing is right. So the current proposal is only acceptable where no physical barrier exists - and therefore the description is wrong. ( And adding footways properly where pedestrians are kept separate from vehicles also requires additional ways. Simply adding yet more tags for something which is not part of the actual roadway is another 'macro' bodge which in reality requires at least a separate way ... ) -- 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] Divided roads proposal
On Sun, Dec 6, 2009 at 4:16 PM, Steve Bennett stevag...@gmail.com wrote: Example: http://www.openstreetmap.org/?lat=-37.820693lon=144.919989zoom=18layers=B000FTF This looks right to you? I think it could be rendered better, especially if there is width information provided, but I think it's mapped right, or at least the best it can be mapped without using areas for everything. That said, if you only limited your proposal to roads like this where there is essentially no routing information contained in the divider data, I won't oppose it. On Sun, Dec 6, 2009 at 3:56 PM, Steve Bennett stevag...@gmail.com wrote: Yahoo — now that I look — does exactly what I'm proposing, rendering a single road with a dotted line down the middle to indicate the division. Yahoo seems to be indicating a center turning lane with that rendering, not a physical divider. They happen to also be mapping those small physical dividers that way, but I suspect that's out of lazy mapping rather than anything else. I think it is essential that slipways are mapped. ESPECIALLY when one is trying to add the right routing instructions. TomTom has started showing motorway and major road slipway details properly. You need to know when to get to an inside lane and take a slip road PRIOR to the actual junction. Maybe we should be mapping slipways, hopefully there's a better approach than marking them all as fully fledged roads though. I think now is a good time to bring up the rule against using the word road on this mailing list. To put it more clearly, ways do not have a one to one correspondence with roads. They never have. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Google Maps maps them as single roads. Not around here. http://tinyurl.com/yz8y8dh Another good example: http://tinyurl.com/ye76u9v I suspect what you're seeing in cases where Google Maps is using a single road is simply Google Maps not getting around to fixing its data yet. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Fri, Dec 4, 2009 at 6:49 AM, Steve Bennett stevag...@gmail.com wrote: I don't know how to convince you that I'm not proposing changing the way major roads and dual carriageways are mapped. This is about minor divisions in minor roads. I see your point but how can you avoid people using this for major roads as weel ? A primary road is minor compared to motorways, so it can be used for primary highways ?. And where is the limit, why should we fix a limit, what is the criteria of using of not using it, the width between the two carriage ways, the height, the length ? This should be clarified in you proposal. Pieren ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On 04/12/2009, at 06.49, Steve Bennett wrote: I don't know how to convince you that I'm not proposing changing the way major roads and dual carriageways are mapped. This is about minor divisions in minor roads. Let me ask you: how do you think that a road with a painted traffic island down the middle should be mapped: 1) As a single road with no special tagging 2) As a single road with a tag indicating the division 3) As two separate ways, with ways connecting them every time there's a gap in the painted division. This discussion is reminiscent of other discussions just like it. There are two orthogonal approaches to mapping in the OSM. One is the drawing approach, which is the most intuitive, since it reminds of the way maps have been drawn with paper and pencil for hundreds of years. The drawing approach is favoured by newbies and people who mostly care about having a beautiful, detailed map to look at. The other approach is the tagging appoach, where details about the landscape are mapped onto a line passing through it. The tagging approach to mapping puts emphasis on a graph representation of the surrounding world which for example can be used for routing, area- and distance determination, and may types of statistical and database applications, situations where the drawing method doesn't work. Therefore, the tagging method has some digital properties that are incridibly useful and powerful, and it is these properties that is the innovative secret behind the incredible success of OSM. The regrettable fact is that more and more mappers don't realize the incredible power of the tagging approach. Drawing dual carriage roads as two separate ways is appealing in many ways since that's how it looks. However, it creates problems in more complicated situations, for example if intersections with other roads, with cycle- and pedenstrian paths are involved, because the connectivity very easily gets screwed up. Keeping the connectivity correct forces you to draw things that aren't representitive of the real world, and suddenly, the drawing isn't pretty. The result is, such as I've seen lately, that people start to say why should we care about routing at all? One example: an intersection of two crossing dual carriageways will result in four nodes. If the intersection is regulated by a traffc light, you will need to specify four traffic lights, but to OSM this will appear as four intersections, and passing through the intersection, your GPS will think you have to pass two intersections instead of one. So, here, while drawing dual carriageways as two ways looks right, it is wrong from a topological point of view. There's nothing you can do with drawing that can't be done with tagging, and there's in principle nothing being done with tagging that can't be rendered beautifully on the final map. It only depends on the richness of the tagging language and the sophistication of the renderers. It is very important that OSM keeps its head straight and doesn't succumb to the increasing pressure of making the map a 2D-drawing of the world. Cheers, Morten ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
2009/12/4 Morten Kjeldgaard m...@bioxray.dk There's nothing you can do with drawing that can't be done with tagging, and there's in principle nothing being done with tagging that can't be rendered beautifully on the final map. It only depends on the richness of the tagging language and the sophistication of the renderers. I just agree partly 1) if you would map intersections of dualcarriage-ways precisely according to the standards, there would be just 1 intersection node at junctions, because at the junction there is no physical barrier and therefore the dualcarriage-way gets 1 way, it doesn't stay 2 parallel ways. 2) there is lots of situations where you can't easily represent complex geometrical configurations just with tagging. Tagging works well for regular shaped stuff, but when it comes to more fluid situations (not classical roads with standard lanes but areas like in pedestrian zones, living_streets, medieval situations, ... you get the best representation by drawing an area. Squares are one of the key features for representation of the individual structure of a city. There shapes tell you about the history and the influencing factors that interacted while the urban tissue evolved. 3) Drawing dual carriage ways as 2 parallel ways is a very easy way to solve lots of junctions, where you would otherwise have to add an awful lot of turning restrictions (or split the way and add divider.tags/relations). 4) It offers also more positional accuracy as well, as dual carriage ways usually consist of more lanes than the other roads. See here: http://maps.google.it/maps?hl=deie=UTF8ll=41.912905,12.481273spn=0.002822,0.004823t=hz=18 situations like this are IMHO not resolveable (with appropriate effort) just by tagging. 5) Keep it simple. Here are lots of experts discussing about the ideal scheme, but there are far more users that don't participate in list discussions (or even read them) and contribute less frequently and IMHO this group will grow much faster then the expert group. The easier it is for them to participate, the more they will become. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal [tagging vs. area mapping]
Morten Kjeldgaard wrote: On 04/12/2009, at 06.49, Steve Bennett wrote: This discussion is reminiscent of other discussions just like it. Indeed it now reminds me of the mapping everything as areas thread that I recently started. There are two orthogonal approaches to mapping in the OSM. One is the drawing approach, which is the most intuitive, since it reminds of the way maps have been drawn with paper and pencil for hundreds of years. The drawing approach is favoured by newbies [..] [..] There's nothing you can do with drawing that can't be done with tagging, and there's in principle nothing being done with tagging that can't be rendered beautifully on the final map. It only depends on the richness of the tagging language and the sophistication of the renderers. It is very important that OSM keeps its head straight and doesn't succumb to the increasing pressure of making the map a 2D-drawing of the world. I'm one of those newbies who got tempted and that was the reason why I asked the questions in the above-mentioned thread. Now I am increasingly realizing the power of connectivity and how the useful graph results from the node way tagging approach that seems to be the foundation of OpenStreetMap as it stands today. Area mapping does have a role, but it now seems to be more specialized than my innocent eyes used to believe. But although I'm learning the ways of OpenStreetMap, I'm learning by observing what others do and experimenting on my own. It works and in due time I'm sure I'll become a decent mapper, but that process could be more efficient. Some people on this list, such as you, seem to have a good grasp of the concepts. Would you be so kind as to start a wiki page that explains good practices and discernment in choosing what should be mapped as a tagged line and what really requires an area ? I'm sure that the whole project would benefit if arguments could be articulated in a demonstration that would convince the wondering newbies such as me. There is surely not a single best way, but the need for a correct connected graph is definitely a good frame into which our methods must fit. Leveraging your experience to get everyone to understand that will surely be a significant contribution to data quality. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
2009/12/4 Martin Koppenhoefer dieterdre...@gmail.com: 2009/12/4 Morten Kjeldgaard m...@bioxray.dk There's nothing you can do with drawing that can't be done with tagging, and there's in principle nothing being done with tagging that can't be rendered beautifully on the final map. It only depends on the richness of the tagging language and the sophistication of the renderers. I just agree partly Me Too, however There are two types of Dual Carrage-way.which I think is the problem here. Type One. Large Motor Way Type Roads (but also large Primary roads), with Slip Roads, and the two separate carriageways are never joined, or maybe by a roundabout from time to time or other major junction. Here its best to map each carriage way separately as they are separate roads. Type Two Smaller roads where the Dual Carriage way exists but is split at regular intervals by gaps to turn right (Sorry I'm in the Uk (Left in most countries)), Often created due to the lay of the land, Bridges, Tunnels, Crossing Islands, Safety Island, etc. May exist for a few meters upto a major junction to help traffic flow etc I think we actually need both, and use the right one for the logic of the road/junction. Peter. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Hi all, I have rewritten the proposal, using the ideas from here. I've gone for maximum simplicity, removing all the single-direction u-turn stuff and the relations on adjoint roads. Please have a read and comment: http://wiki.openstreetmap.org/wiki/Proposed_features/Divided_road Perhaps I should have started a new proposal instead of cannibalising this one so much. If Vovanium objects, I'll roll it back, and move mine to a new page. But it's probably best to just have one discussion. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal [tagging vs. area mapping]
On Fri, Dec 4, 2009 at 9:22 PM, Jean-Marc Liotier j...@liotier.org wrote: I'm sure that the whole project would benefit if arguments could be articulated in a demonstration that would convince the wondering newbies such as me. There is surely not a single best way, but the need for a correct connected graph is definitely a good frame into which our methods must fit. Leveraging your experience to get everyone to understand that will surely be a significant contribution to data quality. I'm new too, and would also appreciate insight. My take in the brief period of time that I've been doing this is that the connected graph is the first priority, and everything else is eye candy. A correctly mapped out, but disconnected, pedestrian mall is much less useful than a simple path that connects to stuff, for example. With the advent of high quality aerial photography, it becomes very tempting to want everything to line up perfectly with the photo. But there are some cases where it's just not possible: - bike paths that connect with roads: since the line is down the centre of the road, and the path, something has to give to make them connect. - building outlines: I think it's more important to give a shape which people will recognise, than perhaps the most accurate tracing around the base of the building. - big open areas like in industrial sites, that nonetheless have clear pathways through them. Mapping the pathways is probably more useful than the vast expanse of asphalt. - areas with lots of little footpaths that connect. At some point, we should filter down and map just the most relevant, important ones. (But as I said, I'm new.) Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
+1 (on tagging vs drawing models) I think it's a good rule of thumb that if it can't be rendered easily, then you need to look again at your tagging model - your data isn't structured in a way that is usable Never tag for a renderer. Always tag for the renderers. Richard On Fri, Dec 4, 2009 at 9:35 AM, Morten Kjeldgaard m...@bioxray.dk wrote: This discussion is reminiscent of other discussions just like it. There are two orthogonal approaches to mapping in the OSM. One is the drawing approach, which is the most intuitive, since it reminds of the way maps have been drawn with paper and pencil for hundreds of years. The drawing approach is favoured by newbies and people who mostly care about having a beautiful, detailed map to look at. The other approach is the tagging appoach, where details about the landscape are mapped onto a line passing through it. The tagging approach to mapping puts emphasis on a graph representation of the surrounding world which for example can be used for routing, area- and distance determination, and may types of statistical and database applications, situations where the drawing method doesn't work. Therefore, the tagging method has some digital properties that are incridibly useful and powerful, and it is these properties that is the innovative secret behind the incredible success of OSM. The regrettable fact is that more and more mappers don't realize the incredible power of the tagging approach. Drawing dual carriage roads as two separate ways is appealing in many ways since that's how it looks. However, it creates problems in more complicated situations, for example if intersections with other roads, with cycle- and pedenstrian paths are involved, because the connectivity very easily gets screwed up. Keeping the connectivity correct forces you to draw things that aren't representitive of the real world, and suddenly, the drawing isn't pretty. The result is, such as I've seen lately, that people start to say why should we care about routing at all? One example: an intersection of two crossing dual carriageways will result in four nodes. If the intersection is regulated by a traffc light, you will need to specify four traffic lights, but to OSM this will appear as four intersections, and passing through the intersection, your GPS will think you have to pass two intersections instead of one. So, here, while drawing dual carriageways as two ways looks right, it is wrong from a topological point of view. There's nothing you can do with drawing that can't be done with tagging, and there's in principle nothing being done with tagging that can't be rendered beautifully on the final map. It only depends on the richness of the tagging language and the sophistication of the renderers. It is very important that OSM keeps its head straight and doesn't succumb to the increasing pressure of making the map a 2D-drawing of the world. Cheers, Morten ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Fri, Dec 4, 2009 at 5:04 PM, John Smith deltafoxtrot...@gmail.comwrote: This opens up a can of worms about micro mapping. You either need to split the way to accomodate your suggestion or some other micromapping technique to accomlish this, in any case you are adding almost nothing that a legal turning restriction won't presently accomplish. I don't think painted lines are the same thing as properly divided ways and don't really need anything that special to distinguish them. Ok, I've uploaded 10 examples of median strips here: http://wiki.openstreetmap.org/wiki/Talk:Proposed_features/Divided_road#Images_for_discussion Hopefully we can use them to discuss whether this proposal is helpful, and what alternatives there might be otherwise. In most of these images, I think either we would still map it with two ways (even if the divided=* tag exists), or we would currently just leave it as a single way, missing out on detail. Would love to hear your thoughts there. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Fri, Dec 4, 2009 at 8:27 AM, Steve Bennett stevag...@gmail.com wrote: Ok, I've uploaded 10 examples of median strips here: http://wiki.openstreetmap.org/wiki/Talk:Proposed_features/Divided_road#Images_for_discussion Hopefully we can use them to discuss whether this proposal is helpful, and what alternatives there might be otherwise. 1, 2. Dual carriageway 3. Assuming these roads eventually divide, I'd personally use a dual carriageway, however, this is disputed. Currently, this could be best handled with three highway areas, with the highway in the center marked as access=no (in jurisdictions where crossing that gore area is prohibited, I can't tell the jurisdiction from the photo). In jurisdictions where those painted gore areas mean use a lot of extra care when crossing, this could be mapped as a single way. 4. Looks easiest to map accurately as a highway area with several barrier=* areas. Are we allowed to make highway multipolygons? Or just ignore it, as it's so insignificant, and use a single way. I don't see any other markings in this photo, if there are signs, I've ignored them because I can't see them. I see no lane markings, so it fits with area=yes, in the context of roads, indicates that the area has no street lines within it. 5, 6, 7, 8, 9, 10. Dual carriageway Regarding 3, hopefully we can eventually get a solution for directed highway areas, whether that means using both an area and a way, or (my current preference) using a relation with role=left, role=right, and role=outer, then oneway=yes to represent that vehicles must travel forward with respect to the left and right ways. However, they would really be unnecessary here, because both the incoming and outgoing ways would already be oneway=yes, so there's basically no way a router would direct you backward on them (I guess it might if you asked for directions from one spot to another within the same area!). ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Peter Childs wrote: There are two types of Dual Carrage-way.which I think is the problem here. Type One. Large Motor Way Type Roads (but also large Primary roads), with Slip Roads, and the two separate carriageways are never joined, or maybe by a roundabout from time to time or other major junction. Here its best to map each carriage way separately as they are separate roads. Type Two Smaller roads where the Dual Carriage way exists but is split at regular intervals by gaps to turn right (Sorry I'm in the Uk (Left in most countries)), Often created due to the lay of the land, Bridges, Tunnels, Crossing Islands, Safety Island, etc. May exist for a few meters upto a major junction to help traffic flow etc I think we actually need both, and use the right one for the logic of the road/junction. +1 There is a very important point here: there's not a single unifying idea that we can come up with that describes every situation. Ultimately, beauty is in simplicity. The example with the motorway is interesting, because on one hand, we think of it as the motorway from A to B, but when analysing its topology, it is perhaps more correct to think of it as one road from B to A and another road from A to B... that happens to be next to it. The distinctive feature is that the two roads are actually not connected directly, i.e. you need to travel along another road in order to reverse directions. The second example you mention is on the other end of the spectrum. A split road is from a topological standpoint a single road, since you could make a U-turn on it (some places even without breaking the law) if you want to reverse directions. I think these are useful rules-of-thumb to keep in mind. In practice, there are other issues, such as the number of T- and X- intersections etc. that may influence how a particular segment of road is best mapped as simply as possible. Jean-Marc Liotier wrote: But although I'm learning the ways of OpenStreetMap, I'm learning by observing what others do and experimenting on my own. It works and in due time I'm sure I'll become a decent mapper, but that process could be more efficient. Some people on this list, such as you, seem to have a good grasp of the concepts. Would you be so kind as to start a wiki page that explains good practices and discernment in choosing what should be mapped as a tagged line and what really requires an area ? In connection with a discussion we are currently having on the talk-dk list about how to map cycleways I recently made a page on the wiki [0]. The purpose of the page is to explore the different problems/benefits you encounter when mapping cycleways as separate ways, vs. mapping them as tagged onto the road itself. These problems are similar to the current discussion on dual carriageways. Making this wiki page was an exercise to enlighten myself and (hopefully) others, and I learned a lot from making these diagrams. The page is in danish, but you can probably grasp the meaning by looking at the diagrams. If there's interest, I can translate the text to english. (You can't use the diagrams as best practice examples, because some of them are actually incorrect or bad practice :-) ) Wrt. the line vs. area problem, it's interesting, but a little bit OT in this thread, so let's discuss it separately (so not to lose focus.) Cheers, Morten [0] http://wiki.openstreetmap.org/wiki/Da:Cykelstier ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Fri, Dec 4, 2009 at 10:03 AM, Morten Kjeldgaard m...@bioxray.dk wrote: In connection with a discussion we are currently having on the talk-dk list about how to map cycleways I recently made a page on the wiki [0]. The purpose of the page is to explore the different problems/benefits you encounter when mapping cycleways as separate ways, vs. mapping them as tagged onto the road itself. These problems are similar to the current discussion on dual carriageways. Yeah, they're pretty much exactly the same. At least, here in Florida. Bicycles are allowed to use a dedicated bike lane, or, in certain situations such as when preparing to make a turn, can use the regular roadway. So using two separate ways doesn't really convey the reality of the situation. We really have a (for example) three lane highway, with two lanes for cars and bicycles, and one lane for bicycles only. There is nothing but a thick white line separating the bike lane from the other lanes, so using more than one way is inappropriate. Basically, we need to find a way to map lanes, and we need a way to connect those lanes along ways and not just at individual nodes. The latest strategy I've been thinking about is to use areas. Areas can be connected together along ways, and not just at individual nodes, so it's almost perfect. Perfect until you try to add a oneway tag. I'll have to draw something up... 1--2 || 3--4 || || 5--6 || 7--8 || || 9--0 way id=34 nd ref=3 nd ref=4 tag k=divider v=line tag k=line v=lane_change_allowed (or v=lane_change_prohibited, or v=change_with_extra_care, etc.) /way relation id=1234 member type=way ref=12 role=left member type=way ref=13 role=outer member type=way ref=34 role=right member type=way ref=24 role=outer tag k=type v=directed_area tag k=highway v=cycleway tag k=oneway v=yes /relation relation id=3456 member type=way ref=34 role=left member type=way ref=35 role=outer member type=way ref=56 role=right member type=way ref=46 role=outer tag k=type v=directed_area tag k=highway v=secondary tag k=oneway v=yes /relation relation id=5678 member type=way ref=56 role=outer member type=way ref=57 role=outer member type=way ref=78 role=outer member type=way ref=68 role=outer tag k=type v=multipolygon tag k=barrier v=traffic_island tag k=surface v=grass tag k=foot v=yes /relation relation id=7890 member type=way ref=90 role=left member type=way ref=79 role=outer member type=way ref=78 role=right member type=way ref=80 role=outer tag k=type v=directed_area tag k=highway v=secondary tag k=oneway v=yes /relation Editor support would obviously be helpful, but isn't an absolute requirement. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Anthony wrote: On Fri, Dec 4, 2009 at 8:27 AM, Steve Bennett stevag...@gmail.com wrote: Ok, I've uploaded 10 examples of median strips here: http://wiki.openstreetmap.org/wiki/Talk:Proposed_features/Divided_road#Images_for_discussion Hopefully we can use them to discuss whether this proposal is helpful, and what alternatives there might be otherwise. 1, 2. Dual carriageway 3. Assuming these roads eventually divide, I'd personally use a dual carriageway, however, this is disputed. Currently, this could be best handled with three highway areas, with the highway in the center marked as access=no (in jurisdictions where crossing that gore area is prohibited, I can't tell the jurisdiction from the photo). In jurisdictions where those painted gore areas mean use a lot of extra care when crossing, this could be mapped as a single way. 4. Looks easiest to map accurately as a highway area with several barrier=* areas. Are we allowed to make highway multipolygons? Or just ignore it, as it's so insignificant, and use a single way. I don't see any other markings in this photo, if there are signs, I've ignored them because I can't see them. I see no lane markings, so it fits with area=yes, in the context of roads, indicates that the area has no street lines within it. 5, 6, 7, 8, 9, 10. Dual carriageway Seconded - I think in most of these cases, the connection to the other roads dictates that they need to be split and correctly join other ways. 5 for example would seem to be best described as a 'roundabout' as it certainly looks like it is for turning around. 3 as actually just a wide road marking. Such cross hatch areas do need to be covered but this a single road with 'advisory' road marking rather than a divider? -- 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] Divided roads proposal
On Fri, Dec 4, 2009 at 11:49 AM, Lester Caine les...@lsces.co.uk wrote: 3 as actually just a wide road marking. Such cross hatch areas do need to be covered but this a single road with 'advisory' road marking rather than a divider? I don't know. I actually got a traffic ticket once for going through a cross hatch area, and I researched it in depth to try to see whether or not it was legit. In the end, I couldn't find any definitive statement one way or the other, and I wound up paying the ticket rather than taking a day off work to try to fight it. I think I was right, that it basically means use lots and lots of extra caution when changing lanes (*), but I wasn't 100% sure (plus I figured there would then be a dispute as to whether or not I used lots and lots of extra caution). (*) A solid white line means lane change is discouraged, wide lines are used for extra emphasis, and chevrons are used for special emphasis. But the lines are referred to as channelizing lines, and the area is referred to as a gore area, which suggests to me that it might be considered legally equivalent to non-roadway. In any case, I think it would be useful to map these painted areas as areas, regardless of whether or not it is technically legal to drive over them (which probably varies from jurisdiction to jurisdiction anyway). If they're illegal to drive over, we can then add access=no, which satisfies people who want routing information for emergency vehicles or for people who just don't care about breaking the law. Unfortunately, connecting an area, to a way which is adjacent to it, doesn't work. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal [tagging vs. area mapping]
2009/12/4 Steve Bennett stevag...@gmail.com I'm new too, and would also appreciate insight. My take in the brief period of time that I've been doing this is that the connected graph is the first priority, and everything else is eye candy. A correctly mapped out, but disconnected, pedestrian mall is much less useful than a simple path that connects to stuff, for example. depends what you want to do with the map. When focusing on routing outside the mall: yes, when interested in routing for far distances: doesn't matter, when interested in routing inside the mall you'll prefer to have details and won't care whether it's connected to the outside... - bike paths that connect with roads: since the line is down the centre of the road, and the path, something has to give to make them connect. yep, that's a small problem - building outlines: I think it's more important to give a shape which people will recognise, than perhaps the most accurate tracing around the base of the building. IMHO the better the outline is matching reality the better you will recognise the building (but we could have several outlines for different heights, I agree, e.g. a tower-building/skyscraper often has a bigger base and gets taller due to zoning-laws) - big open areas like in industrial sites, that nonetheless have clear pathways through them. Mapping the pathways is probably more useful than the vast expanse of asphalt. I do generally both. - areas with lots of little footpaths that connect. At some point, we should filter down and map just the most relevant, important ones. everyone is free to map whatever he/she likes. As long as you don't delete footpaths mapped by other people because you don't consider them relevant, that's not a problem. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
2009/12/3 Steve Bennett stevag...@gmail.com: I find the current practice of duplicating minor roads when there is a median strip pretty unsatisfactory. Even disregarding the effort, the end result never renders well: usually the street name is written twice, the one-way arrows get messy etc. That's a problem with the rendering, not with the mapping. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 9:39 AM, John Smith deltafoxtrot...@gmail.comwrote: 2009/12/3 Steve Bennett stevag...@gmail.com: I find the current practice of duplicating minor roads when there is a median strip pretty unsatisfactory. Even disregarding the effort, the end result never renders well: usually the street name is written twice, the one-way arrows get messy etc. That's a problem with the rendering, not with the mapping. Unless you want to write routines for pre-processing two almost-parallel ways back into a single way so it can be rendered neatly, I suggest it's a mapping problem. Don't make work for other people if you don't have to. Everything about a dual carriageway can (now) be expressed using turn restriction relations, and it would probably be better if it were done that way. Richard ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 11:17 AM, Richard Mann That's a problem with the rendering, not with the mapping. Unless you want to write routines for pre-processing two almost-parallel ways back into a single way so it can be rendered neatly, I suggest it's a mapping problem. Don't make work for other people if you don't have to. It's both, a rendering issue when the names are duplicated and a mapping issue about representing the turn restrictions. Note that It's a very old definition in OSM to say that we duplicate the ways when there is a physical division between the two ways (it's physically impossible to cross the road with a car, e.g. a sidewalk, a wall, a fence, grassland). A simple painted white line is not enough. I personnaly prefer to draw two ways and the connections where they exist than creating dozen turn restrictions. It's funny to see the whole discussions about drawing every single lanes or everything as areas and this proposal. Pieren ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 9:17 PM, Richard Mann richard.mann.westoxf...@googlemail.com wrote: That's a problem with the rendering, not with the mapping. Unless you want to write routines for pre-processing two almost-parallel ways back into a single way so it can be rendered neatly, I suggest it's a mapping problem. Yep, a very hard rendering problem, or an easy mapping problem. Everything about a dual carriageway can (now) be expressed using turn restriction relations, and it would probably be better if it were done that way. I don't want to be as ambitious as genuine dual carriage motorways, but it would be great for minor roads. Here's an example I came across (I didn't map it): http://www.openstreetmap.org/?lat=-37.824316lon=144.940393zoom=18layers=B000FTF See Hartley St there? It's a tiny pathetic service road that happens to have some concrete blocks down the middle of some sections. I can just picture how nicely this would render with the right tagging and styles... Another example, nearby: http://www.openstreetmap.org/?lat=-37.81993lon=144.960348zoom=18layers=B000FTF Queensbridge St, where it splits from the bridge itself. The two lanes are actually separated by a decent chunk of concrete. The northbound lane then splits again just before hitting Flinders St. I don't blame the mapper for not representing all this detail, because ultimately it doesn't really make a difference to anyone using the map, and it will render as a mess. Another really cool thing you could do with this proposal (not currently suggested) is a width parameter. Then you could really precisely control the separation of the lanes, which are now a bit hit and miss - it's very hard to specify how much gap there is, because you have to guess how wide the renderer will think the road is, then compensate for that. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Sorry, one last example, also nearby: http://www.openstreetmap.org/?lat=-37.826282lon=144.947554zoom=18layers=B000FTF This mess might be much more understandable if pairs of lanes that were physically together were rendered as pairs (with a line between them), and those that were on separate physical carriageways had the separation as at present. Note that It's a very old definition in OSM to say that we duplicate the ways when there is a physical division between the two ways Is that an argument against changing it? Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On 03/12/2009, at 10.39, John Smith wrote: 2009/12/3 Steve Bennett stevag...@gmail.com: I find the current practice of duplicating minor roads when there is a median strip pretty unsatisfactory. Even disregarding the effort, the end result never renders well: usually the street name is written twice, the one-way arrows get messy etc. That's a problem with the rendering, not with the mapping. And that's a problem with a knee-jerk reaction to a quite reasonable discussion. -- Morten ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Ok, I've mocked up what it might look like. Complicated to show you though: 1) http://www.geowiki.com/halcyon/ 2) Lat: -37.821995 Lon: 144.919573 3) Add these lines to the of end of the big edit box: way[highway=service][!divider] { color: white; width: 3; casing-width: 5; } way[divider=median][highway=service] { z-index:5; color:white; width:7; casing-width:9; linecap:round;} { z-index:6; color:brown; width:2;} Pretty, and a huge improvement on real separated lanes. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
2009/12/3 Steve Bennett stevag...@gmail.com Can I draw some attention to this: http://wiki.openstreetmap.org/wiki/Proposed_features/Divided_road I was going to propose exactly the same thing, pretty much. Maybe I'd quibble with some of the naming. I am also working on a proposal (let's say it's in an very early phase) to cope with dividers but also the opposite (no divider but linear possibility to change way/lane), so I'd consider this in a more generalized way. My first draft is here, you're welcome to comment: http://wiki.openstreetmap.org/wiki/User:Dieterdreist/drafts it allows to define things like: there is a kerb between the footway and the street, but on given nodes there is a lowered kerb to crossover. Or: these two ways are separated by a grass-divider (what does actually permit pedestrians to cross at any place) Or: this link is connected from the motorway between Node A and Node B (probably would need to be a way, so splitting required) Or: this is a lane that will turn right on next crossing (highway=lane, turn=right), and changing to this lane is possible from A to B, from B to C there is a uninterupted line, while from C to D the way must be mapped separately (as it is divided physically). This would require somehow double mapping and tagging (as there has to be a artificial way that connects the original (main) way to node C, (which is the last changeover possibility for the lane) as a normal way (so that applications can do routing considering the lanes but will work also not knowing about them). This should also work to map pedestrian areas beside roads: between this road and this building there is a pedestrian area. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 10:15 PM, Martin Koppenhoefer dieterdre...@gmail.com wrote: it allows to define things like: there is a kerb between the footway and the street, but on given nodes there is a lowered kerb to crossover. Yes, it's certainly quite expressive, at the cost of complexity - and having to use relations. One of the advantages of simply setting a divider= tag is its simplicity and succinctness. One tag, and you've got something that will render nicely, and be useful for routing. I see four essential tag values: divider=none (can freely cross) divider=marked or something (do not route a car across this, but a pedestrian could cross it) divider=median_strip (again, do not route a car across this, but you could drive across it in an emergency) divider=barrier (a physical barrier that would probably stop a pedestrian) Not certain about the key or tag names. Maybe: divided=no divided=line divided=median divided=barrier Does that work? median=? I guess we also have to be careful about the word barrier. Setting a highway to be barrier=yes would normally mean the highway itself is somehow impassable to people not on it...but wouldn't say anything about u-turns. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Oops, I also meant to point out a screenshot of my mockup: http://wiki.openstreetmap.org/wiki/Image:Stevage_Divided_road.png (And an earlier version: http://wiki.openstreetmap.org/images/archive/9/9e/20091203113817!Stevage_Divided_road.png ) Sorry for the spam. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 12:47 PM, Steve Bennett stevag...@gmail.com wrote: http://wiki.openstreetmap.org/wiki/Image:Stevage_Divided_road.png If dividers are so small and do not create any turn restrictions, I just ignore them. But you are lucky to have a source which allow you to go so deep in details. Anyway, your rendered example looks nice but you create other difficulties like drawing the street name. Pieren ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
2009/12/3 Steve Bennett stevag...@gmail.com On Thu, Dec 3, 2009 at 10:15 PM, Martin Koppenhoefer dieterdre...@gmail.com wrote: it allows to define things like: there is a kerb between the footway and the street, but on given nodes there is a lowered kerb to crossover. Yes, it's certainly quite expressive, at the cost of complexity - and having to use relations. +1 One of the advantages of simply setting a divider= tag is its simplicity and succinctness. One tag, and you've got something that will render nicely, and be useful for routing. but haven't we already barrier for this? I see four essential tag values: divider=none (can freely cross) divider=marked or something (do not route a car across this, but a pedestrian could cross it) divider=median_strip (again, do not route a car across this, but you could drive across it in an emergency) divider=barrier (a physical barrier that would probably stop a pedestrian) yes, I see these as possible dividers/barriers for my proposed area-relation. The area-relation allows optionally to add a divider/barrier-role, which can be either area (e.g. if it is a very thick irregular wall), way or node (nodes can be barriers, but also the opposite: entrances (interruptions, openings), gates, lift_gates, lowered_kerbs, ...) And it can solve the lane-problem (map lanes explicitly and do not confound them with other highway-ways), connect bicycle-tracks to the adjacent road and define the height of the kerb (without having to map the kerb, but beeing able to map the kerb-lowerings), cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
moved to proposal from user-space to http://wiki.openstreetmap.org/wiki/Relations/Proposed/Area cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
It's both, a rendering issue when the names are duplicated and a mapping issue about representing the turn restrictions. Note that It's a very old definition in OSM to say that we duplicate the ways when there is a physical division between the two ways (it's physically impossible to cross the road with a car, e.g. a sidewalk, a wall, a fence, grassland). A simple painted white line is not enough. Agreed, while it's sensible for two ways in a dual carriageway, it seems OTT to have two ways for a road simply because it's got white lines down the middle. It seems to me to be introducing unnecessary complication, not to mention increasing required storage size and bandwidth for transmitting the data or the need for pre-processing tools to remove it if you're not interested in it for your purposes. Nick ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 4:06 AM, Steve Bennett stevag...@gmail.com wrote: I find the current practice of duplicating minor roads when there is a median strip pretty unsatisfactory. I was thinking about this recently when we had the map everything as areas thread, and I have to agree with you to some extent (though not completely). Duplicating and separating makes more sense when the roads are completely separate - separate entrance/exit ramps, etc. For a road with a median strip, we're better off mapping the median strip on top of the road. I'm not sure how this would work without using areas, though. And even then, it'll be complicated. I think the proposal at http://wiki.openstreetmap.org/wiki/Proposed_features/Divided_road is far too kludgy and temporary to justify going through the database and merging ways which were separated to represent roads with traffic islands. And doing so will likely lose data - specifically, the size of the divider (which varies often enough to make divider:width=* an insufficient solution). On Thu, Dec 3, 2009 at 5:17 AM, Richard Mann richard.mann.westoxf...@googlemail.com wrote: On Thu, Dec 3, 2009 at 9:39 AM, John Smith deltafoxtrot...@gmail.com wrote: 2009/12/3 Steve Bennett stevag...@gmail.com: I find the current practice of duplicating minor roads when there is a median strip pretty unsatisfactory. Even disregarding the effort, the end result never renders well: usually the street name is written twice, the one-way arrows get messy etc. That's a problem with the rendering, not with the mapping. Unless you want to write routines for pre-processing two almost-parallel ways back into a single way so it can be rendered neatly, I suggest it's a mapping problem. Don't make work for other people if you don't have to. The thing is, you kind of have to. There has never been a one-to-one correspondence between roads and ways. It's not just dual carriageways that sometimes get rendered poorly, it happens when ways are split as well. Mappers could assist the renderers by creating relations, so every single way which is part of Dale Mabry Highway (as in http://en.wikipedia.org/wiki/Dale_Mabry_Highway) gets put into a single relation. I think that's more the way to go. Solves lots of problems in addition to that one. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Fri, Dec 4, 2009 at 12:40 AM, Nick Whitelegg nick.whitel...@solent.ac.uk wrote: Agreed, while it's sensible for two ways in a dual carriageway, it seems OTT to have two ways for a road simply because it's got white lines down the middle. It seems to me to be introducing unnecessary complication, not to mention increasing required storage size and bandwidth for transmitting the data or the need for pre-processing tools to remove it if you're not interested in it for your purposes. Ok, so no one has raised objections so far. What's the next stage to getting this adopted? Open the RFC? Work on the proposal itself more first? For example I'm not sure I see the need for the relation - but maybe I'm not understanding it. As I see it, when there's a junction: | ==+== | Either the N-S street can go through, or it can't. That's a simple flag best set on the node (junction=*). By default I'd say it doesn't go through, it's two separate side streets each feeding into a street with a median strip. Incidentally, I've uploaded a pic showing three types of divider: http://wiki.openstreetmap.org/wiki/Proposed_features/Divided_road I was just mapping out a new suburb, and really realising how ridiculous it is to make two lanes every time there's a divided street. I really hope we can get this implemented. It would also be useful at roundabouts, where the streets are briefly divided when they meet the roundabout. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Fri, Dec 4, 2009 at 2:40 AM, Anthony o...@inbox.org wrote: I'm not sure how this would work without using areas, though. And even then, it'll be complicated. I think the proposal at http://wiki.openstreetmap.org/wiki/Proposed_features/Divided_road is far too kludgy and temporary What do you find kludgy about it? I thought it was quite elegant? It makes the division a property of the road, and lends itself to assisting with routing. to justify going through the database and merging ways which were separated to represent roads with traffic islands. I don't think anyone was proposing doing that. I mean, one *could* but I don't see what the compulsion would be. And doing so will likely lose data - specifically, the size of the divider (which varies often enough to make divider:width=* an insufficient solution). Well, a couple of comments there: 1) Do you think the divider widths as currently recorded are particularly accurate? 2) Do you think it's important? 3) Do you think in a situation where we had accurate information about a non-uniform-width divided road, we would even consider throwing out that information and replacing it with this? Of course not. The point is, certainly in the areas I'm working in, there are lots of back streets with median strips that are sketched out fairly quickly, and there's nothing particularly interesting or important about the width of that strip, and using two separate ways to indicate the lanes on either side - *that's* the kludge. It's implying a higher level of accuracy and detail than we have. The thing is, you kind of have to. There has never been a one-to-one correspondence between roads and ways. It's not just dual carriageways that sometimes get rendered poorly, it happens when ways are split as well. Yep, this proposal doesn't solve every problem. Mappers could assist the renderers by creating relations, so every single way which is part of Dale Mabry Highway (as in http://en.wikipedia.org/wiki/Dale_Mabry_Highway) gets put into a single relation. I think that's more the way to go. Solves lots of problems in addition to that one. Yeah...but before that happens, the interface for editing relations has to be really natural. And it would still be much more work than simply marking a stretch of road. Better rendering is just one benefit. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 10:58 AM, Steve Bennett stevag...@gmail.com wrote: On Fri, Dec 4, 2009 at 2:40 AM, Anthony o...@inbox.org wrote: I'm not sure how this would work without using areas, though. And even then, it'll be complicated. I think the proposal at http://wiki.openstreetmap.org/wiki/Proposed_features/Divided_road is far too kludgy and temporary What do you find kludgy about it? I thought it was quite elegant? It makes the division a property of the road, and lends itself to assisting with routing. It doesn't seem to be general enough, but instead as an incomplete lump of special cases. to justify going through the database and merging ways which were separated to represent roads with traffic islands. I don't think anyone was proposing doing that. I mean, one *could* but I don't see what the compulsion would be. So routers are going to have to handle two completely different ways of doing the same thing? And doing so will likely lose data - specifically, the size of the divider (which varies often enough to make divider:width=* an insufficient solution). Well, a couple of comments there: 1) Do you think the divider widths as currently recorded are particularly accurate? The ones I've done are. 2) Do you think it's important? Not of huge importance, but I think it's something important enough that it should eventually be recorded. 3) Do you think in a situation where we had accurate information about a non-uniform-width divided road, we would even consider throwing out that information and replacing it with this? Of course not. See above regarding mapping the same thing two different ways. The point is, certainly in the areas I'm working in, there are lots of back streets with median strips that are sketched out fairly quickly, and there's nothing particularly interesting or important about the width of that strip, and using two separate ways to indicate the lanes on either side - *that's* the kludge. It's implying a higher level of accuracy and detail than we have. I guess I'm spoiled with access to Yahoo aerials. Still, if there's nothing interesting about the median, why bother mapping it in the first place? You can still add the turn restrictions, which is the important part. Mappers could assist the renderers by creating relations, so every single way which is part of Dale Mabry Highway (as in http://en.wikipedia.org/wiki/Dale_Mabry_Highway) gets put into a single relation. I think that's more the way to go. Solves lots of problems in addition to that one. Yeah...but before that happens, the interface for editing relations has to be really natural. I don't see why it does. In fact, I think you're putting the cart before the horse. Define the relation in the wiki, and the editors will follow. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Fri, Dec 4, 2009 at 3:14 AM, Anthony o...@inbox.org wrote: It doesn't seem to be general enough, but instead as an incomplete lump of special cases. It covers the small, but salient, case of divided roads. Nothing more. We're just talking about one key here. So routers are going to have to handle two completely different ways of doing the same thing? I'm not sure what you're getting at. Topologically speaking, the way a divided road is handled now is the same as two parallel streets with occasional cross streets. Routers handle that case in the way they handle any other street routing. With the new divider tag, a divided road is handled just like a single street - again, a case it already has to handle. The only difference is that it has to check before turning right (or left in othe countries), when u-turning, and when crossing a junction. All these checks already have to be performed, and there's just a very small amount of additional logic. (I haven't seen router code, I'm just speculating here, but the tests involved seem trivial.) 1) Do you think the divider widths as currently recorded are particularly accurate? The ones I've done are. 2) Do you think it's important? Not of huge importance, but I think it's something important enough that it should eventually be recorded. I should also have asked whether the widths of the roads are accurately measured. Presumably we need to know the width of each road and the distance between them, at each point. But if you want to model the roads that accurately, probably best to keep them as separate lanes, or use an area tag or something. I guess I'm spoiled with access to Yahoo aerials. Still, if there's nothing interesting about the median, why bother mapping it in the first place? You can still add the turn restrictions, which is the important part. Why? Primarily as a landmark, I would think. The fact that there is a median strip is more important, relatively, than that there is a median strip which is 83cm wide, surfaced in terracotta pavers overgrown with moss... Anyway, to back up slightly here, the benefits of this proposal are: - much simpler and faster entry of minor divided roads - better rendering at little cost - cleaner data structure that better reflects what we're trying to map The downsides are: - less specificity of the width of the division Are there other downsides I'm missing? Could you explain the lack of generality argument - what cases can't it handle, and why is this important? Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 12:05 PM, Steve Bennett stevag...@gmail.com wrote: I should also have asked whether the widths of the roads are accurately measured. Presumably we need to know the width of each road and the distance between them, at each point. But if you want to model the roads that accurately, probably best to keep them as separate lanes, or use an area tag or something. And I think that's eventually where we're going. The distance between the centerlines is only part of the equation, but I wouldn't want to throw that information away. This is all moot, however, because I now understand that you have no intention of throwing that information away. The old way of doing things would still be allowed, and in fact would not be deprecated, right? Are there other downsides I'm missing? I think the biggest downside is that it creates two accepted ways to map the same thing. Even that, I suppose, is not a problem, if we make it clear that the old way, which contains more information, is preferable. Your initial email suggested to me that this method of mapping was superior to the old one. You made a statement, which I agree with, that the current practice of duplicating minor roads when there is a median strip [is] pretty unsatisfactory. So I thought you were attempting to replace that current practice. You've clarified that you're not, and it looks like a router/renderer would be safe ignoring these new tags, so I guess I don't object. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 12:05 PM, Steve Bennett stevag...@gmail.com wrote: Are there other downsides I'm missing? Oh yeah, I forgot to mention a big problem. Using forward/backward breaks when a way is reversed. So divider=u_turn_forward and divider=u_turn_backward are a bad idea. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Steve Bennett wrote: Why? Primarily as a landmark, I would think. The fact that there is a median strip is more important, relatively, than that there is a median strip which is 83cm wide, surfaced in terracotta pavers overgrown with moss... Anyway, to back up slightly here, the benefits of this proposal are: - much simpler and faster entry of minor divided roads - better rendering at little cost - cleaner data structure that better reflects what we're trying to map The downsides are: - less specificity of the width of the division Are there other downsides I'm missing? Could you explain the lack of generality argument - what cases can't it handle, and why is this important? The basic decision has to be at what level you switch between macro and micro mapping. If the feature is so small, is there any need to describe it in a complex series of tags ( assuming lots of gaps down a short road )? As soon as you move to a mapping zoom that can display he fine detail, then an area detail is required anyway. One of the examples provided was a simple 'U' - with the detail provided, one assumes a continuous dividing structure, with one way movement around the whole length. If there are gaps in the divide, then those should be indicated, with links or additional tags, but if the divide is just the odd bollard then it would be better drawn as a single dead end road anyway? This is more than adequate for routing purposes, and any finer detail really needs the areas of all of the parts that make up the roadway? At what point to you switch from 'dual carriageway' to 'divided road' ? -- 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] Divided roads proposal
On Fri, Dec 4, 2009 at 4:21 AM, Anthony o...@inbox.org wrote: And I think that's eventually where we're going. The distance between the centerlines is only part of the equation, but I wouldn't want to throw that information away. This is all moot, however, because I now understand that you have no intention of throwing that information away. The old way of doing things would still be allowed, and in fact would not be deprecated, right? Oh, of course. Sorry, I hadn't made that clear. :) In most cases, the trade-off between extreme accuracy, and usability will lead to the simpler divider=* tag. For more complex cases, or where the shape of the individual lanes matters, I'd still make two individual ways. I think the biggest downside is that it creates two accepted ways to map the same thing. Even that, I suppose, is not a problem, if we make it clear that the old way, which contains more information, is preferable. I think it depends on the circumstance, and whether that information is valuable. If the community feels strongly that duplicated roads are better in most cases, then the divided tag will be marginalised to only the smallest cases, like painted traffic islands. I suspect it will become more popular than that, though. Your initial email suggested to me that this method of mapping was superior to the old one. You made a statement, which I agree with, that the current practice of duplicating minor roads when there is a median strip [is] pretty unsatisfactory. So I thought you were attempting to replace that current practice. Well, put it this way: if this was implemented, I would duplicate far fewer roads in future. I don't know if I would go around converting existing ones, though. I do think it's superior for minor roads where the division is a modern traffic-calming device. You've clarified that you're not, and it looks like a router/renderer would be safe ignoring these new tags, so I guess I don't object. If a router ignores the tags, it will occasionally try and send you through a barrier. If a renderer ignores them, then it might be hiding important information from the consumer. But there are certainly already plenty of divided roads which are not marked as such. Oh yeah, I forgot to mention a big problem. Using forward/backward breaks when a way is reversed. So divider=u_turn_forward and divider=u_turn_backward are a bad idea. Yeah I haven't really thought through the u_turn thing much. I think it's awkward for the reason you describe. It's probably better to just have a junction tag that indicates there is a physical gap, and use turn restrictions if you don't want one direction using it. Or just create two lanes, and put a one-way in the right spot. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
2009/12/3 Anthony o...@inbox.org Are there other downsides I'm missing? I think the biggest downside is that it creates two accepted ways to map the same thing. Even that, I suppose, is not a problem, if we make it clear that the old way, which contains more information, is preferable. no, I think this is a big con: there are (and will always be) people who change the map to different schemes, sometimes also loosing information (e.g. on talk-de it was recently reported that people delete separately mapped cycleways and attach cycleway=track to the road, what is loosing information (positional and inhebitance of tagging different maxspeeds, surfaces, widths, etc.). If the new way is containing less information, don't set it online. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 12:37 PM, Martin Koppenhoefer dieterdre...@gmail.com wrote: 2009/12/3 Anthony o...@inbox.org Are there other downsides I'm missing? I think the biggest downside is that it creates two accepted ways to map the same thing. Even that, I suppose, is not a problem, if we make it clear that the old way, which contains more information, is preferable. no, I think this is a big con: there are (and will always be) people who change the map to different schemes, sometimes also loosing information (e.g. on talk-de it was recently reported that people delete separately mapped cycleways and attach cycleway=track to the road, what is loosing information (positional and inhebitance of tagging different maxspeeds, surfaces, widths, etc.). If the new way is containing less information, don't set it online. That was my initial reaction. But then I thought about all the other situations where we have two accepted ways to map the same thing. For example, a building can be mapped as a point or an area. One is more detailed, but the other is acceptable if you don't have time for the detail. I don't know. I read over the proposal again and I don't even get it, actually. Is the way supposed to be split before and after each intersection? ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 12:38 PM, Steve Bennett stevag...@gmail.com wrote: Well, put it this way: if this was implemented, I would duplicate far fewer roads in future. If it worked (and I really haven't delved into the details enough to check), I might be convinced to add division information where I otherwise wouldn't bother (because the current method of creating a dual carriageway is a pain). I don't know if I would go around converting existing ones, though. If you don't knowthat scares me. If a router ignores the tags, it will occasionally try and send you through a barrier. For a U-turn? Whether or not a U-turn is legal should probably be treated as unknown by routers in most cases. We're currently not really mapping that at all. OTOH, if you're talking about a left turn, I thought the point was that this would be handled by turn restrictions just like would be done currently. If a renderer ignores them, then it might be hiding important information from the consumer. But there are certainly already plenty of divided roads which are not marked as such. Yeah, I see this as maybe a potential solution in between using a single unmarked way and using a dual carriageway. Although, looking at the proposal, it doesn't look finished - something you seem to admit yourself. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Fri, Dec 4, 2009 at 4:46 AM, Anthony o...@inbox.org wrote: I don't know. I read over the proposal again and I don't even get it, actually. Is the way supposed to be split before and after each intersection? Maybe I should write up the proposal as I see it, but all I'm proposing is: divider=* tag on ways that are divided, where * is probably one of line, median, barrier. junction=*something*, to indicate a gap in a divided road where another road passes through it. ie: | =+= | With junction=fourway or whatever at the +. You could equally well implement the junction like this: | =*+*= | Where * is a node, + is a junction, - is a normal way, and = is a way marked with divider=*. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Fri, Dec 4, 2009 at 4:54 AM, Anthony o...@inbox.org wrote: On Thu, Dec 3, 2009 at 12:38 PM, Steve Bennett stevag...@gmail.com wrote: Well, put it this way: if this was implemented, I would duplicate far fewer roads in future. If it worked (and I really haven't delved into the details enough to check), I might be convinced to add division information where I otherwise wouldn't bother (because the current method of creating a dual carriageway is a pain). Exactly. If a router ignores the tags, it will occasionally try and send you through a barrier. For a U-turn? Whether or not a U-turn is legal should probably be treated as unknown by routers in most cases. We're currently not really mapping that at all. I guess that's country specific. Here (Australia), u-turns are legal by default. But I'm not sure why routing software would need to do u-turns in general. OTOH, if you're talking about a left turn, I thought the point was that this would be handled by turn restrictions just like would be done currently. I'll assume you're in a right-hand-side drive country? Let's take this situation: B | A==C You're going from A to B. I'm saying that the router doesn't attempt to make the turn because the way A-C is marked as divided=median, and there's nothing at the intersection to overrule that. If left turns are possible there, then the junction should be marked as a gap, or there should be a section of non-divided at that point. Turn restrictions are a pain to edit, so it's great that the router gets a free hint here. Yeah, I see this as maybe a potential solution in between using a single unmarked way and using a dual carriageway. Although, looking at the proposal, it doesn't look finished - something you seem to admit yourself. Yep, I didn't write that proposal, I just liked the ideas in it. I guess tomorow I'll write up the summary of this discussion. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
(Weird, did this email not get sent before - so many emails going back and forth. Oops.) On Fri, Dec 4, 2009 at 4:37 AM, Martin Koppenhoefer dieterdre...@gmail.com wrote: no, I think this is a big con: there are (and will always be) people who change the map to different schemes, sometimes also loosing information (e.g. on talk-de it was recently reported that people delete separately mapped cycleways and attach cycleway=track to the road, what is loosing information (positional and inhebitance of tagging different maxspeeds, surfaces, widths, etc.). If the new way is containing less information, don't set it online. Well, I guess I agree with you, but that's a big if. And it's a question of mapping practice, not of software functionality. There are a few scenarios: 1) There is a divided road that has been carefully mapped out. Converting this to a simple divided=* tag could lose information. 2) There is a divided road that has been sketched out roughly, simply to indicate the division. (Very common, I think) Converting this to a simple divided=* tag doesn't lose information, and better indicates the actual level of information stored. 3) There is a divided road which is represented as a single road, because it's not particularly important, or the mapper was lazy or whatever. This could be a *gain* of information. As an example of 3), see this randomly chosen spot: http://www.openstreetmap.org/edit?lat=-37.747252lon=144.7879zoom=19 (Turn on nearmap to see the imagery). There's a corner with a brief division as a safety device. You probably wouldn't map that as two lanes, because it would look ugly and superfluous. But it would look elegant with a simple divided=* tag. Here's one that looks like 2): http://www.openstreetmap.org/edit?lat=-37.77825lon=144.830416zoom=19 There are two lanes to indicate the divided highway, but the mapper hasn't marked out the cut throughs, so this results in worse mapping than if they'd just made a single road. - the router can't tell that you can turn left across the median strip. If you follow that road south a bit, you'll find cases of 3 and 2: painted lines, some of which are marked as separate lanes, some aren't. IMHO it would be cleaner to represent all of those with a divided=* tag, but tthat's just my opinion. http://www.openstreetmap.org/edit?lat=-37.77888lon=144.833565zoom=19 A clear example of 1: obviously you would not convert this massive median strip into a single way. http://www.openstreetmap.org/edit?lat=-37.786651lon=144.84961zoom=19 Another 1), I think. An interesting shaped median strip best left as two ways. http://www.openstreetmap.org/edit?lat=-37.785579lon=144.854427zoom=19 But just a bit down the road, the median strip is very uniform. It's probably wide enough to justify the two lanes, but it's worth asking the question. http://www.openstreetmap.org/edit?lat=-37.791938lon=144.86358zoom=20 Also worth asking whether a traffic island is different from a short division. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 12:58 PM, Steve Bennett stevag...@gmail.com wrote: ie: | =+= | With junction=fourway or whatever at the +. fourway would be the only tag that's not ambiguous. Your junction was already solved properly by turn restrictions. (http://wiki.openstreetmap.org/wiki/Relation:restriction) divider=none, divider=legal, and divider=physical is about the only part of that proposal that I'm fairly sure would work. And divider=physical goes against current mapping principles. As a sort of todo tag, I guess it isn't horrible, though. divider=physical meaning this should be a dual carriageway, but I don't really feel like doing all the work of mapping it that way. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Thu, Dec 3, 2009 at 1:05 PM, Steve Bennett stevag...@gmail.com wrote: Let's take this situation: B | A==C You're going from A to B. I'm saying that the router doesn't attempt to make the turn because the way A-C is marked as divided=median, and there's nothing at the intersection to overrule that. If left turns are possible there, then the junction should be marked as a gap, or there should be a section of non-divided at that point. So if I decide to split AC for whatever reason, I've just broken the routing? Relations aren't that hard. Not nearly hard enough to justify something like this. IMHO. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
2009/12/3 Richard Mann richard.mann.westoxf...@googlemail.com: Unless you want to write routines for pre-processing two almost-parallel ways back into a single way so it can be rendered neatly, I suggest it's a mapping problem. Don't make work for other people if you don't have to. I thought we were trying to map the world in as much detail as possible, just because they are almost parallel doesn't mean they don't diverge etc. I like the current way of doing things where if they are physically seperated they should be drawn as 2 different ways, some even argue a strong centreline is enough but I disagree on this point. Also I was under the impression that turning restrictions are a signage thing, not a physical barrier thing. If the map doesn't render the data properly then it's a rendering problem, otherwise we're just throwing away data because someone thinks it will look nicer that way. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
Awesome, nice to finally get some proper criticism :) I'll write up the proposal as I see it. I think most of these comments are assuming that somehow a single divider=* tag is going to replace all split roads, including dual carriageway motorways etc. That's not what I'm suggesting. I like the current way of doing things where if they are physically seperated they should be drawn as 2 different ways, some even argue a strong centreline is enough but I disagree on this point. There's a good example of why a divider=* tag makes sense. Currently you have two options: single road, or two separate roads. This is a good middle ground for cases where the two separate roads strategy is too big and heavy. Also I was under the impression that turning restrictions are a signage thing, not a physical barrier thing. Not sure what you mean here. We're discussing three different types of divider: those that are just legal/signed/marked, those where there is a physical kerb but which could theoretically be driven over, and those with a serious barrier that even a pedestrian would struggle with. If the map doesn't render the data properly then it's a rendering problem, otherwise we're just throwing away data because someone thinks it will look nicer that way. Again, this proposal is not primarily about rendering. That's a nice benefit in some cases. The goals are: 1) More appropriate data structure 2) Better usability 3) Better rendering at no effort. There may be other benefits too. And again, I'm mostly focused on the situations where mapping as a single road is losing information, and mapping as two roads is inappropriate/undesirable. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
2009/12/4 Steve Bennett stevag...@gmail.com: Again, this proposal is not primarily about rendering. That's a nice benefit in some cases. The goals are: 1) More appropriate data structure How is this more appropriate, you are loosing real world information to improve rendering. 2) Better usability How does this improve useability at all? 3) Better rendering at no effort. Number 3 shouldn't be a goal, just like using the layer tag shouldn't be used to alter the rendering order. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
On Fri, Dec 4, 2009 at 4:28 PM, John Smith deltafoxtrot...@gmail.comwrote: 2009/12/4 Steve Bennett stevag...@gmail.com: Again, this proposal is not primarily about rendering. That's a nice benefit in some cases. The goals are: 1) More appropriate data structure How is this more appropriate, you are loosing real world information to improve rendering. The main focus is *adding* information to roads currently mapped as single (non-divided) roads. Gaining information. Not losing. There are a huge number of places that this will add information that was not previously mapped. I don't know how to convince you that I'm not proposing changing the way major roads and dual carriageways are mapped. This is about minor divisions in minor roads. Let me ask you: how do you think that a road with a painted traffic island down the middle should be mapped: 1) As a single road with no special tagging 2) As a single road with a tag indicating the division 3) As two separate ways, with ways connecting them every time there's a gap in the painted division. Here's an example that we can use to discuss (use satellite view): *http://tinyurl.com/y8cxujs ** *If you would like to propose some rules for the use of this tagging (like don't use it for dual carriageways), that would be helpful. 2) Better usability How does this improve useability at all? Because adding a single tag is a lot simpler than making two parallel ways. Because modelling one road as is easier than modelling one road as two roads. 3) Better rendering at no effort. Number 3 shouldn't be a goal, just like using the layer tag shouldn't be used to alter the rendering order. Better rendering should always be a goal. What I think you mean is we shouldn't modify the way we map in order to compensate for some short-term deficiency in current renderers. I agree. I think modelling minor streets that have median strips as single roads is the right thing to do. Better rendering is just an additional benefit. If there was a trade-off between them, we would go with correct modelling. Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Divided roads proposal
2009/12/4 Steve Bennett stevag...@gmail.com: The main focus is *adding* information to roads currently mapped as single (non-divided) roads. Gaining information. Not losing. There are a huge number of places that this will add information that was not previously mapped. This opens up a can of worms about micro mapping. You either need to split the way to accomodate your suggestion or some other micromapping technique to accomlish this, in any case you are adding almost nothing that a legal turning restriction won't presently accomplish. I don't think painted lines are the same thing as properly divided ways and don't really need anything that special to distinguish them. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk