On Thu, Aug 27, 2009 at 2:06 PM, John Smith<delta_foxt...@yahoo.com> wrote: > > Besides why should you care about needing this explicit information, if it's > rendered you will see a sign, you will also see the nearest junction and your > mind can put 2 and 2 together. A computer can do the exact same thing.
You're asking why should we tag things explicitly? Because we're building a database. A huge, complex database that's used by lots of different people and software all over the world. And as I've said before, fudging a solution always seems like a great idea at the time (e.g. "oh, this'll do the job, it's easier, and it's good enough - why bother doing it properly/explicitly") - until it breaks due to unforeseen circumstances. > This is why I keep saying this is tagging for software, you are explicitly > tagging for software to "know" which junction the stop sign applies to, where > as just like you it can "see" a junction and it can "see" a stop sign and it > will "know" that the stop sign applies to that junction. > > So if we can't tag for rendering we aren't allowed to tag for routing > software either. Please listen to me. A requirement to stop *intrinsically involves a way AND an intersection*. A requirement to stop **IS** an interaction between a way AND an intersection. This is why I would suggest using a relation. Not because a relation is easier for software, but because a relation describes the **nature** of the thing to be described. But hey, I'll go along with the majority in whatever's decided. _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk