[OSM-talk] Bing maps is misplaced

[Steve Bennett]

On Sun, Dec 19, 2010 at 11:04 PM, M∡rtin Koppenhoefer
<dieterdreist at gmail.com> wrote:
> This discussion is simply about the quality level: are you satisfied
> with probable information derived from an aerial photo depicting the
> situation some years ago, or do you want to insert only information
> you verified on the ground and you can guarantee for?

Ah, cool. For me, this is a no-brainer: very comfortable with
"probable" information. I'd rather have 1000 streets at 90% accuracy
than 10 streets at 100% accuracy. Yes, that means I've created many
errors in OSM.

Btw, no idea how a ground survey would give a better idea of
highway=tertiary vs residential. Also, all suburban streets (of which
the example was clearly one) are access=yes, no question there.

John F Eldredge wrote:
>So, you are saying that you feel OpenStreetMap should reflect the status of the road when the aerial photo was made, >rather than the current status?

That's a rather unkind question, implying a non-existent choice: that
you could either map every road from the air, or map every road from
the ground. It also implies, incorrectly, that mapping from the ground
ensures that OSM is always up to date. Ground-mapped data is only as
current as the most recent visit, and I'd wager that the number of
people regularly checking existing data with ground surveys is a
fraction of the number using ground surveys to add new data. Whereas
by contrast with something like NearMap (updated multiple times per
year for large cities) I'm frequently looking over areas with existing
data.

I would argue that it's precisely because we use aerial imagery so
much that OSM data for Melbourne (Australia) is fairly up to date. I
can recall several instances where I mapped a housing development that
had already been partially added by someone else - that is, OSM has
actually tracked the building of the individual streets. I can't
imagine that happening with ground surveys.

Steve