<html><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /></head><body style='font-size: 10pt; font-family: Verdana,Geneva,sans-serif'>
<p> </p>
<div> </div>
<p>On 2015-11-24 13:20, Martin Koppenhoefer wrote:</p>
<blockquote type="cite" style="padding: 0 0.4em; border-left: #1010ff 2px solid; margin: 0">
<div dir="ltr">
<div class="gmail_extra"><br />
<div class="gmail_quote">2015-11-24 12:43 GMT+01:00 Colin Smale <span><<a href="mailto:colin.smale@xs4all.nl">colin.smale@xs4all.nl</a>></span>:<br />
<blockquote class="gmail_quote" style="margin: 0px 0px 0px 0.8ex; border-left: 1px solid #cccccc; padding-left: 1ex;">
<p>One issue with dual carriageways (and now I think about it, also railway lines) is about generalisations at certain zoom levels. If you zoom out beyond a certain level, both halves of a DC (or the individual tracks of a railway) would be better modelled as a single line. The renderer/consumer needs an algorithm, and/or hints from the tagging, to know what belongs to what.</p>
</blockquote>
<div>yes, generalization would be nice sometimes for zoomed out levels, but a relation would not help much. You still have to judge which part can be omitted and which not (or how to combine the two (or more) into one), at least for situations where both parts are not strictly parallel.<br /><br /></div>
<div>See here how it works not too bad like it is done now:<br /><a href="http://www.openstreetmap.org/#map=16/42.1322/13.1322">http://www.openstreetmap.org/#map=16/42.1322/13.1322</a><br /><a href="http://www.openstreetmap.org/#map=15/42.1314/13.1323">http://www.openstreetmap.org/#map=15/42.1314/13.1323</a></div>
</div>
</div>
</div>
</blockquote>
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote">
<div>I see how the casing on the lines gets suppressed where they overlap, but you can see the name being rendered individually on the two halves. And I suspect the shields with the road numbers are more numerous than they need to be. That casing-suppression looks a bit wierd sometimes where two roads are geometrically adjacent WITHOUT being logically related, for example here (the north-south road right in the middle of the map, where it is labelled "Canterbury Way"). Two parallel but unrelated roads, of different classes, are being blended into one:</div>
<div> </div>
<div>http://www.openstreetmap.org/#map=15/51.4347/0.2426</div>
<div> </div>
</div>
</div>
</div>
<blockquote type="cite" style="padding: 0 0.4em; border-left: #1010ff 2px solid; margin: 0">
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote">
<div> </div>
<blockquote class="gmail_quote" style="margin: 0px 0px 0px 0.8ex; border-left: 1px solid #cccccc; padding-left: 1ex;">
<p>They can provide certainty where otherwise complex heuristics would be needed which may or may not work in all cases. Two streets with the same name isn't good enough on its own IMHO. The two halves of a dual carriageway may have different names and still be the same road.</p>
</blockquote>
<div> </div>
<div>Given that there is no such unambiguous thing in the real world like "the same road", you'd always have to judge based on a common definition. There will likely be arguments for both ways of looking at it: same road or different roads. A relation could make it clear how the mapper saw it, but I'd already question the concept on a more general level: what is "the same road"?</div>
<div> </div>
</div>
</div>
</div>
</blockquote>
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote">
<div>
<div>The definition of "the same road" will probably depend on the specific use case. The use case may be rendering, to produce a nice-looking map on the screen or on paper, or it <strong>might</strong> be navigation-related. My idea is to remove the constituents of the relation and replace them with a simplified version, which looks/behaves more suitably for the given zoom level. Keeping all the roads properly connected will be a helluva job, so it might not be viable as a simplification of a routing network, but giving the resulting turn directions based on the simplified model may have some benefits; not sure about this though.</div>
</div>
<div> </div>
<div> </div>
<div>//colin</div>
</div>
</div>
</div>
</body></html>