[OSM-talk] Road Widths
Thomas Walraet
thomas at walraet.com
Thu Jan 11 03:15:58 GMT 2007
Ben Robbins a écrit :
>
> I'm talking about all the roads that are just unclassified. Lanes=0.5
> doesnt seem logical. I'm specifically talking about the 4 variations of
> unclassified roads that can be found relative to the width of the
> average veicle.
I like the "lanes" option :
- It's self-explanatory (don't have to look on the wiki what 'type3' mean)
- It could be used directly by renderers (width proportional to this
factor, don't draw road under a certain factor for a given scale, etc)
- If someone wants to use 3 variations, he can. If someone wants to use
8 variations, he can.
The only problem is just the choice of the key value : lanes. I think
this tag should be kept to store the number of lanes. (really)
Here we talk about storing the road width relative to the width of the
average vehicle (as you mention). We could use k="carwidth" for example...
Now, why I'm not fond of highwaytype :
- It mix informations that could be interesting to separate : road width
and side composition.
- It's limited to 4 types _you_ need. I think someone will add type5 for
road so small that a large car should be careful. Then type0 for road
large enough for normal vehicles going in each direction to drive past
each other normally even if there is cycle on the side. And so on...
- We need to remember what each type definition. And everyone using this
tag should agreed on what each type definition.
- Renderers will need to have rules for each types.
- Your type definitions loose all sense for oneway roads.
More information about the talk
mailing list