Disclaimer: I'm new to the OSM mailing lists, so hopefully I didn't<br>
screw up my email :D<br><br>~~~~~~~~~<br><br>I must say that I find both "Relation:type=stop" and Roy's ideas<br>interesting. The former suggests that we abstract reality away while<br>the latter suggests we mirror reality in OSM.<br>
<br>Mirroring reality will always result in more precision, but there is<br>often a tradeoff for simplicity. For instance, OSM maps multi-lane<br>roads with a single way. Although mapping individual lanes would<br>result in a map closer to reality, there are few benefits and the<br>
work required for mapping, rendering, and routing would be<br>significant.<br><br>Unless there are a number of benefits that arise from knowing the<br>precice location of a stop sign it seems more reasonable to use<br>relations for this job. They're slightly easier to map (no precision<br>
surveying required) and far easier to get software working with. The<br>one downside I can see is that you loose the ability to know exactly<br>WHERE a sign is -- information that I'm hard-pressed to think of a<br>use for (not to say that there isn't a use, mind you).<br>