[OSM-talk] State of the Map slides

[David Earl]

On 27/07/2010 00:23, Dave F. wrote:
> On 25/07/2010 19:56, David Earl wrote:
>> My two talks for State of the Map
>> * Tag Central - a schema for OpenStreetMap
>> * What I learned making a real map on real paper for real people and
>> real money
>> are now available online at http://www.frankieandshadow.com/sotm10/
>>
>
> It's disappointing that in the 'real map' document you twice falsify the
> data to suit your niche rendering:
> Page 18, 26.


That's why I included the slide, goes to the heart of what the talk was 
getting at, and why page 26 is in a section labelled 'Truth'. So I hold 
up my hand an say yes, it is a lie, in my opinion.

But that was one of the things I learned - clients *want* you to lie 
sometimes. In the purist world of OSM where nearly all map renderings so 
far are driven by us not paying clients, we have so far had the choice. 
In the real world, we are going to hit this kind of thing more and more 
often. The constraints are different when you have a client calling the 
shots, and truth is open to interpretation. There were other areas on 
the map I mentioned where I "lied" too, sometimes by omission, but I 
worked around that locally without touching the data.

I argued repeatedly that the footpath case was misleading, but in the 
end they are the client. It's not as clear cut as maybe I have suggested 
though. There is indeed a public footpath (right of way) running over 
that land, it is signposted on the ground (albeit rather obscurely). For 
its first few metres the right of way, the footpath, is part of the 
access to a pub car park.

I should probably change it back on the OSM data now. What I really need 
to do is to provide a way for my renderer to lie. I already allowed the 
renderer to, for example, adjust positions of captions related to OSM 
ids, and I guess I could and should have done that for this case too, to 
say for this ID the map should render the service road as a footway, and 
not change the original data.

The one that troubled me more, actually, and is more fundamental, was 
the one on p27. This was complicated, but it turned out that the formal 
order allows cycling (which is what the map shows), but the No Entry 
signs on the ground prohibit it, and apparently No Entry signs are 
illegal to cycle past even if they are wrong. Truth here is a very mixed 
up affair.

Not sure what you mean by page 18 - that's an acknowledgements page. 
Perhaps you mean page 19, the "roundabout". Are you making the same 
judgement about this feature from the abstract impression you are 
getting from the map or do you actually know the junction in question?

Is this junction a roundabout: http://osm.org/go/0EFYMXav2-- ?

Firstly, like a lot of roundabouts, it has traffic lights on it. Is a 
one way system with traffic lights a roundabout, or some different kind 
of junction? Secondly it has a short cut across it. Does that make it 
not a roundabout? Or is part of it a roundabout, and if so is it the 
inner or outer part that is included in the roundabout? Does a 
roundabout have to be round? If so this does not qualify, 
notwithstanding the shortcut, because it is oval. But if a roundabout 
does not have to be round, then in principle this junction (in 
Cambridge) qualifies as a roundabout - http://osm.org/go/0EQSLOVlH-- - 
as it has all the other characteristics of a roundabout, though I doubt 
many locals including me would consider it as such.

I mention those examples because the topology of the Wisbech roundabout 
- http://osm.org/go/0ERWt0NCV-- - is exactly the same as the Stansted 
one and in fact the Cambridge one. It is more convoluted, because of the 
way it crosses the river, but essentially it works to route traffic the 
same and when you see it on the ground, you would see that was the 
traffic engineer's intent.

The sticking point here is, of course, whether they should be treated as 
two separate junctions and if so whether the western one is a 
roundabout. Does a roundabout actually have to form a complete ring 
(circular or not) to classify. In every respect other than this it 
*looks* like a roundabout on the ground.

So I think this raises three issues:

1. Does the judgement of someone surveying on the ground carry more 
weight than someone looking at an abstract map?

2. if it is so obvious what something is from the map, why bother 
marking it up in the first place?

3. What are we trying to convey with junction=roundabout in the first 
place? What's it for?

> I can find no reference to 'junction=approach' in the wiki. What does
> this signify?

Frederick is right, I _used_ it to suppress one way arrows. But I used a 
more general tag to identify this very common kind of arrangement as it 
represents a mapping "idiom" commonly found in OSM where a road divides 
into one-way "splays" entering and leaving a roundabout (or other 
junction), like the Google StreetView picture above.

David