[OSM-talk] Road cores and casings on standard Mapnik rendering

Ben Robbins ben_robbins_ at hotmail.com
Sat Jan 21 05:05:48 GMT 2012




>Can you give an example of a junction that doesn't look good to you?
The z13-z18 links I previously gave seem not all to open, so instead here is a cropped z18 sections:http://wiki.openstreetmap.org/wiki/File:Z18crop.pngIt best shows the common problems, coming from the core-casing width issue.

Below is a junction which I made as neat as I could but didn't touch the highway= tags.  
This is how it was added, and how all the junctions I've seen have been around Dubai 
tend to be mapped, and most of the world, where I’ve edited.  At the bottom left there 
is an example of the issue of there not being a service_link tag, or just that the 
motorway_link renders early.  However I do see your point, and if the _link tags were 
later there would be other issues gained for those dealt with.

http://wiki.openstreetmap.org/w/images/2/21/Comparison_-_Junction1.png

Now if we say that a road only moves 'up' in status when it joins a higher status road 
and not before, then I end up with this: (I have now started to map for the renderer, 
although I can see justification in this in reality, where link has to meet it’s higher 
ranking ‘parent’ before it becomes that status.)

http://wiki.openstreetmap.org/w/images/0/07/Comparison_-_Junction2.png
An issue here is the lack of an unclassified_link road, so in the top left the road sits 
on top of motorway_link.If I say that all departing roads must also drop to the status of the road they are 
linking where that road is lower status, and have no actual _link roads then I get this:

http://wiki.openstreetmap.org/w/images/8/8f/Comparison_-_Junction3.png

If I go back 1 step, but remove all 'link' status tags, I get this:

http://wiki.openstreetmap.org/w/images/1/19/Comparison_-_Junction4.png

So, this really brings up 3 separate considerations, and the rendering is a small 
fraction of it, and I'm going to try to explain what’s in my head in a clear way... here goes!


If all road statuses don't
increase before meeting a higher road status; and all road 
status's decrease,
when leaving another road, in preparation for there end-connection 
road status,
then it works on the usage of no ‘links’. 
If links are used, then all status's 
must have a link variant. This is
currently not so.

 

In the event of this not
being possible, due to ‘reality’ dictating road status, and 
therefore going
against the aforementioned criteria so as to ‘map how it is’, then 
‘links’
would again have to be available for all
variants.  However in the event of 
a
none-‘link’ road coming off a ‘link’ road of higher status, it would need to 
devolve to a none ‘link’ road, which would be messy.

 

However common practice
has the road status as increasing to meet the 
fore-coming road status where it
is higher, and holds onto former road status, 
again where it is higher.

 

So the factors causing
this are:

 

Standard editing practice – Having roads promoted
     to the higher of its options.Missing _link values for the smaller road
     types.Different Rendering Widths.

 

This also brushes on ‘map
how it is, not how it renders’, but then we should also 
‘map so it may render
how it is.’  And here in lies the clash.



>One of my favorite renderings is TopOSM: http://toposm.com/us/index.html

>Its rules are very consistent and I like its
progression of road widths.



Nice, haven't seen this before, thanks for that.

 

Cheers,

Ben

 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20120121/4ecb3bec/attachment-0001.html>


More information about the talk mailing list