[OSM-talk] Not attaching polygons to roads

Dave F. davefox at madasafish.com
Wed Feb 26 11:11:52 UTC 2014


On 26/02/2014 10:27, Pieren wrote:
> On Wed, Feb 26, 2014 at 1:33 AM, Clifford Snow <clifford at snowandsnow.us> wrote:
>>> 5m,10m... and there is no reason to virtualy extend them and falsify the
>>> real world.
>> +1
> omg. In the "real world", a highway is not a thin polyline...

Yes, that why he's saying don't attach it to the centreline in the 
belief it's the edge.

>
>> Wouldn't it be nice if the editors wouldn't allow polygon to connect to
>> highways. I understand that it hard to implement, but when you need to make
>> changes to either object, it is a pain.
> Think that, in some parts of the world, you don't have high res.
> images and you cannot count the amout of lanes or see the shoulders or
> the limit between the road and next landuse.

If there's no imagery to work from & the person has never seen the area 
before then they really shouldn't be mapping it. If a visual survey has 
been performed a bit of 'guesstimating' the distance to the centreline 
from the boundary edge is *still* more accurate than attaching it the 
the way of the road.

>   Or that the person adding
> landuse is working at a "municipality" level, not at a "fence" level.
> It was always like this in OSM, the first contributor adds a node for
> a townhall, the second draws or import the building footprint and move
> the tags from the node to the way, etc. The crowd is not mapping at
> the same map scale. Assuming you start to map "fences" and "walls",
> you have to adapt the existing data to your level of contributions.
Agree

> But you shouldn't forbid other contributions if they are not at your
> expectations.

Not sure how you 'forbid' data, but if the contributions are to existing 
data & reduce the information or accuracy, then they should be reverted.

> Again, it's an iterative concept. New contributions increase the
> quality. What is not acceptable is that new contributions decrease the
> level of the mapping (excepted in some cases I could develop)

Unfortunately, that's what's happened in my case.

---
This email is free from viruses and malware because avast! Antivirus protection is active.
http://www.avast.com




More information about the talk mailing list