[OSM-talk] Pedestrian crossings and barriers

Andy Allan gravitystorm at gmail.com
Sun Jul 29 11:11:28 BST 2007


On 7/29/07, Tom Chance <tom at acrewoods.net> wrote:
> If you're so lost that you
> need to know it's a toucan rather than a pelican crossing I submit that
> you should have read the map better in the first place =)

I disagree. Take the following place:
http://www.gravitystorm.co.uk/osm/?lat=6704522.69283&lon=-18328.80133&zoom=15&layers=B00

Heading west on route 37, it looks at first like you have to do a left
onto Latchmere Road, and then a right onto a cycleway. Many people
would dislike doing an immediate right turn across what is potentially
(and actually is!) a busy primary road. However, there is a toucan
crossing out of sight from where you approach Latchmere road. If you
know it's there, you can assume that the cycle route is actually along
the pavement on Latchmere road, and if you know that, you can spot the
faint paint on the pavement. If it was marked as a generic crossing,
that wouldn't help, and if it wasn't a toucan crossing, you could
assume it was irrelevant. (not that I would render them on a cycle
map...)

If you load the relevant part of the data into josm, you'll see that
I've put in a cycleway representing the lane on the pavement (so a
routing algorithm might be able to help you), but the cycle map
doesn't show in enough detail.

The exact same situation happens a few hundred metres further west,
when crossing Falcon Road - there's a hidden toucan crossing out of
sight, and if you know it's there you know to look for the on-pavement
cycle lanes to get to it.

> One
> question to resolve is the difference between traffic_signals and
> crossing, since I've been using the former for pelican crossings and the
> latter for zebra crossings.

I thought traffic_signals were for vehicular junctions - i.e. both
pelican and zebra would fall under highway=crossing, but two roads
meeting at lights would be traffic_signals. Again, something that
needs clarification on the wiki?

Cheers,
Andy




More information about the talk mailing list