[OSM-newbies] highway=traffic_signals
James Ewen
ve6srv at gmail.com
Fri Apr 9 03:33:13 BST 2010
On Thu, Apr 8, 2010 at 1:36 PM, Dave F. <davefox at madasafish.com> wrote:
> I disagree with this. I think they should be accurately mapped where
> they actually are.
>
> This is an example I corrected yesterday:
> http://osm.org/go/eukjJQRVR--
One could argue that your traffic signal placement is inaccurate.
The signals on North Parade Road (687898571) are placed right on the
way, but in reality they are on the left (north) side of the road.
Similarly, the signals on southbound Pulteney Road (687898570) are on
the way, rather than on the left (east) side of the road.
The lights for northbound Pulteney Road (687898568) are in the middle
of the street, so we can play along with your placement.
What about the set of light underneath the railway overpass on the
east side of the roadway? You never placed a node for them! (BTW, who
is supposed to look at those lights?)
As Richard indicated previously, the map is an abstract representation
of the real world. Placing a single node at the intersection of the
ways with the traffic signal associated with it tells us that this
intersection is controlled by traffic signals.
If you're going to micro map, then place the signals at the exact
location of the poles, but then you should also be mapping the edges
of the road, and not just a single line down the center of the
roadway. You would have to map the edges of the sidewalk as well, to
define its location. You couldn't just put a point down for a garbage
can, but you would have to define a square or circle of the proper
size so that it renders properly.
When creating a map, we really need to decide if we are creating an
abstract representation of the world, or a photo-realistic
representation.
This is probably the biggest downfall of the open concept of OSM. Each
mapper has a different idea of what the OSM map should look like. I
think there should be a single node at the intersection defining the
traffic signals, and you think there should be three nodes depicting
the four traffic signals. We can play editing war till the cows come
home, and no one wins.
Unless there's a uniform agreement, it ends up that the user is the
one that loses.
Have look at your example (http://osm.org/go/eukjJQRVR--), and then
look at this example (http://osm.org/go/eu9dRo8@J--)
Based on looking at the first intersection that you have mapped, I
would assume that the lights on Rotherham Road in the second example
would be controlling the intersection with Cliff Hill, where in fact
the light on Rotherham is a pedestrian controlled light for a
crosswalk, and there are no lights at Cliff Hill.
What happens in commercial mapping, is that someone (or committee)
comes up with a set of rules for the employees to follow (such as how
to map a set of lights at an intersection), and then the employees
follow the rule. If they deviate enough, they get shown the door. With
OSM, we don't have a strictly defined set of rules. We kind of have a
very loosely described set of suggestions. Because everyone is a
volunteer, all we can do is ask everyone to try and play nice, and do
their best.
I can not force anyone to change what they are doing, or how they are
mapping. I have no more power or influence than any one else. The best
that I or anyone else can do, is enter into a discussion such as this
one now. From this discussion, we might come to a consensus of how we
should be mapping specific items.
The biggest downfall of the OSM project is also it's largest asset...
the community. The huge number of contributors are leveraged to spread
the work out, and make the workload lower for each individual. This
same huge pool of contributors each bring their own set of ideas of
how things should be done. As a result, uniformity of the data
representation suffers.
Just my ever so humble opinion...
James
VE6SRV
More information about the newbies
mailing list