[OSM-talk] Featured images for christmas week
David James
david at djames.org.uk
Fri Dec 21 22:44:27 GMT 2007
On Fri, December 21, 2007 6:49 pm, Andy Robinson \(blackadder\) wrote:
> David James wrote:
>
>>
>> One question - what's the source for the residential=building data (the
>> rectangle-ish areas at the sides of the roads)?
>>
>
> It's a mixture of manual survey and artistic impression. The beginning
> and end of each run of houses is easy to do by taking a photo of the
> boundary position which when synchronised in JOSM to the GPX track gives
> you a pretty accurate location.
Thank you, that's roughly what I'd guessed (though I'd thought of
waypoints rather than photographs - I must get around to trying out
photographs).
> The setback from the road is guesswork and
> doesn't yet sit right with the width of the road at every zoom level. The
> thickness again is guesswork and where I'm trying to give an impression
> rather than being precise.
It was the thickness that I was wondering about. I suppose without the
high resoilution aerial mapping being available, it's difficult to do
better than inspired guesswork.
> The same applies to the total run of the
> building, which in most cases is actually semi detached or detached
> housing which should really be drawn as individual blocks. Maybe in a
> couple of year's time I'll feel like going to that extent! for now I'm
> just making a single continuous run where there are only residential homes
> along the run. If something else interjects I take a photo of it and tag
> appropriately.
Would this description of your technique be a useful addition to the one
of the Wiki entries on mapping techniques?
>
> One the house number rendering works you will see I hope little numbers
> at the start and end of each run of building. These are set on nodes
> within each building area. I'm also recording the house number opposite a
> side junction.
>
> I would say though that all of this is only really possible if you have
> fully mapped out all the roads and other paths and tracks. I was surprised
> how easy it is to gather the extra data. Doing it on foot, that's the
> only reliable way of ascertaining the house number on each end of a block.
I'm still at the mapping roads level (doing it by car). I figure paths and
tracks come after that, that'll probably need doing on foot - the exercise
will probably do me good.
I have started thinking about mapping the divisions between shops in the
centre of town ...
--
David James
More information about the talk
mailing list