2014-08-20 11:11 GMT+01:00 Ilpo Järvinen ilpo.jarvi...@helsinki.fi:
On Wed, 20 Aug 2014, Will Phillips wrote:
On 20/08/2014 00:02, Martin Koppenhoefer wrote:
Il giorno 19/ago/2014, alle ore 23:45, Will Phillips wp4...@gmail.com
ha
scritto:
I find that by far the most time consuming part of surveying house
numbers is actually adding the data afterwards and for this reason I
think we should be trying to make the tagging quick and
straightforward for mappers wherever possible.
Which editor are you using?
I've always used JOSM.
Speaking to other mappers, they usually agree that data input for addresses
takes 2-3 times as long as the actual surveying.
My experience is that it takes even longer than that especially if the
area contains small houses rather than bigger buildings. Drawing them
all is very slow compared with inputting addr data. I'd say 10-20x.
There are various reasons for this, some specific to mapping where I
live (England).
Why it takes a long time -
1. The usual problems of reconciling the surveyed data with existing data and
the aerial imagery.
2. Adding other detail at the same time. In particular, adding buildings. It
would be much quicker if I chose just to add nodes.
So true.
With the experience of tens of thousands addresses survey, this is really
big obstable. Some of my surveyed data rotted(!) because the drawing delay
hindered immediate input. That is, I lost the near-term memory about
details before I could draw all the building I could easily survey
addresses for.
This is also the reason I really get almost angry when people oppose
importing building without addresses (because they find them useless
without other details such as addresses included already during the
import). I doubt that such people have much experience on surveying
addresses and trying to draw the buildings while inputting the addr data
to OSM.
I agree with these observations from your experience. Back in January
I said I suspected building outlines were unimportant*. But that was
before I started heavily address-mapping! My opinion now is that a
two-step process works really well: first pass gets basic building
outlines (from aerials or maybe even from imports), then second pass
is address mapping on the ground, which will probably also include
lots of changes to the building outlines. But the first pass makes the
second pass a lot easier - both the surveying and the data entry.
Best
Dan
* http://www.openstreetmap.org/user/mcld/diary/20663
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb
