<div class="gmail_quote">On Fri, Dec 4, 2009 at 9:22 PM, Jean-Marc Liotier <span dir="ltr"><<a href="mailto:jm@liotier.org">jm@liotier.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
I'm sure that the whole project would benefit if arguments could be<br>
articulated in a demonstration that would convince the wondering newbies<br>
such as me. There is surely not a single best way, but the need for a<br>
correct connected graph is definitely a good frame into which our<br>
methods must fit. Leveraging your experience to get everyone to<br>
understand that will surely be a significant contribution to data quality.<br>
<br></blockquote><div>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.<br>
<br>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:<br>- 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.<br>
- 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.<br>- 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.<br>
- areas with lots of little footpaths that connect. At some point, we should filter down and map just the most relevant, important ones.<br><br>(But as I said, I'm new.)<br><br>Steve<br></div></div><br>