On Thu, Jul 23, 2009 at 9:48 AM, Steve Hill<st...@nexusuk.org> wrote: > On Wed, 22 Jul 2009, Nicholas Barnes wrote: > >> Should, for example, the component ways making up the roundabout be >> grouped in their own "I'm a roundabout" relationship? > > Do we need to be able to tell which ways are part of a roundabout anyway? > I mean, on the ground a roundabout is just a one-way circular road with > some other roads coming off it - there isn't really anything special about > it that makes it a roundabout. > > The only use of an explicit "I'm a roundabout" tag/relation that I can > immediately think of is to make driving directions more human-readable > (i.e. "At roundabout, take the 3rd exit"). In this case it may be better > for the data user to use some heuristic, much as we do ourselves when we > look at a piece of road. e.g. If there is a closed (clockwise in > the UK) one-way loop with a diameter of less than X metres then consider > it a roundabout when generating human-readable driving directions. > > Using this kind of heuristic would also have the advantage of setting an > application-specific upper bound to the size of a roundabout - when > roundabouts get beyond a certain size then it is probably better for > sat-navs to go back to the usual "take the next left" driving directions > instead of "take the 7th exit".
I disagree - I've come across town-centre one-way systems that are smaller than some large out-of-town roundabouts. There is a clear distinction between them in the way they are signed - e.g. using a "roundabout ahead" warning triangle, so we should in fact record which are roundabouts and which are just circular oneway streets, since they are in fact different on the ground. Cheers, Andy _______________________________________________ Talk-GB mailing list Talk-GB@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-gb
