On 04/14/2013 06:32 AM, Steve Bennett wrote:
Hi,
My view (I'll try to be concise).
Being able to map both abstractions (like a schematic route) and
physical details is a real problem. We need to be able to do both. The
problem is not unique to rail. Use cases I've thought of:
- roads (the road network, vs the individual bits of tarmac)
- rail (the line vs the bits of track)
- power (the power grid vs every individual power line)
- traffic lights ("this intersection has traffic lights" vs each
individual physical traffic light)
- universities, hospitals, precincts (the campus as a whole, rather
than the individual plots of land near each other)
- bike parking (space for 20 bikes here vs 10 individual bike hoops)
- car parking (space for 200 cars here vs several individual parking areas)
- bike routes (the route follows the river, vs the two individual
tracks on each bank)
The point is: it's hard to make beautiful maps without mapping the
abstractions. The physical detail looks ok at high zoom levels, but
when you're zoomed out, it's messy - and it's really not easy to
automatically generate these kinds of abstractions.
It would be really good to have a single, consistent approach
(including terminology) for this "multiple levels of abstraction"
problem.
Yes, completely agreed! I was trying to achieve this in a more limited
sense, with just the transport network: both highways and waterways
already have tagging schemes described that broach this:
- highway vs. area:highway
- waterway=river vs. waterway=riverbank
I consider my proposal as being a logical extension of this line of
thought to railways, where "railway" represents the line, "area:railway"
represents the area occupied by the line, and "railway:track" represents
the individual tracks.
It is unfortunate that highways and waterways are already inconsistent
as described above: in retrospect, area:waterway=river would've been
more consistent (but area:highway remains just a proposal, of course.)
I was considering railways to be the "odd one out" as far as the
transport network is concerned, but the rest of your list are all good
examples of this problem outside of transport that I hadn't yet
considered, though given how hard it is to even get buy-in for a revamp
of railway tagging at this point I'm certainly not encouraged to try to
define a universal mechanism for separating levels of abstraction across
*all* features! :)
_______________________________________________
Tagging mailing list
Tagging@openstreetmap.org
http://lists.openstreetmap.org/listinfo/tagging
