[Openstreetmap-dev] REST
Stefan Holst
mail at s-holst.de
Wed Jul 20 13:59:39 BST 2005
Nick Whitelegg wrote:
> Stefan,
>
> Looks interesting. A couple of points:
>
> - notice each street is represented as a separate track on your schema.
> Would it not make more sense to have each street as a separate track
> *segment* (as I have done)? That way, if you survey a route in your GPS as
> a track, then divide it up into different segments (different roads, paths
> etc) you could upload the whole track back into your GPS if needed?
the problem is, that many roads around here cannot be modelled with a
single tracksegment. i already have two roads which have a mainroad and
sideroad part. some sideroads here also have more than two endings.
i think it makes sense to modell roads as tracks with one or many
segments.
> Motorway (Autobahn, Autoroute, Interstate, etc) : foot 1, horse 1, bike 1,
> car 5
:)
since there are many known integer, your system can (should?) be
refined like this:
type foot - horse - bike - car
perm. footpath 1..9 0 0 0 ( depends on the owner ;) )
footpath 10 0 0 0
combined foot
and bikepath 10 0 1..9 0 ( bikers have to pay attention )
... seperate
trails 10 0 10 0
bridleway 1..9 10 1..9 0
"anlieger frei" 1..9 1..9 1..9 9
minor road 1..9 1..9 1..9 10
... with walkway 10 1..9 1..9 10
main through
road 1..19 1..19 1..19 20
... with sep.
bike and foot
ways 20 1..19 20 20
...
the actual value of the given ranges are determined by
the usage of a road. e.g. seldom used minor roads would
have 9 9 9 10 - busy minor roads would have 1 1 1 10.
clear criterias have to be defined to get an impartial
classification of roadsegments. each national street
system can be used as guidance in a way i sketched
above.
> So if I want all surfaced roads I can select those segments with a "car"
> value of at least 2.
the surface character should be stored seperately since
it is independent of the permission or usage. moreover
the surface is the same for all users.
RY
Stefan
More information about the dev
mailing list