[talk-au] Missing streets in Sydney

Ian Sergeant inas66+osm at gmail.com
Tue Sep 6 03:32:48 BST 2011


I wrote:


> > Personally, I think people shouldn't  map areas when they don't have any
> > knowledge of the topology and layout because I think fixing errors takes
> > several orders of magnitude longer than the tracing.
>

On 6 September 2011 10:31, Stephen Hope <slhope at gmail.com> wrote:

Who cares?


Um, Me?  Maybe you?


> That's not the right comparison. Does fixing the errors
> take several times longer than editing it from nothing?


I think it does, yes. And in the meantime you have data that may be wrong.
Personally, on the local scale where most OSM Australia mapping is up to, I
struggle to see much benefit of having the vectors traced from imagery on a
map over the raw imagery itself.


> The local visit can certainly add more
> details, (street names for a start) but it's damn rare that it will
> get a better overall layout positioning. Lot's of gps traces can be
> just as good, but that takes multiple visits on different days to get
> good averaging.
>

I think this is a proposition that could probably do with more empirical
evidence.  Especially with the halcyon days of nearmap behind us.  I do have
roads and cycleways that I have many many GPS traces for the single way over
multiple visits, and the divergence appears very limited.  Of course GPS
signals can lose it entirely occasionally with reflections, etc, but it
isn't like the surveyor doesn't have the imagery as another arrow in their
quiver.  If things get displaced it isn't hard to highlight areas of
possible concern and investigate the errors further.  If after understanding
the topology, on-the-ground changes, and any offset, the easiest way is to
trace the imagery for a way, then that is an option still open to the
surveyor.

> Any perceived time saving is illusory, when someone has to visit the area
sooner or later
> anyway.

 Again, who cares?


You are only responding to half an argument.  Steve was saying that he can
map 10x the area from imagery than I can from surveying in the same time,
and advocating that as a benefit of imagery tracing.  My response to that
was that the time saving is illusory, because after he maps from the
imagery, I still have to go there and survey it.   Well, words to that
effect anyway...


Beside which, you're wrong.  I've done a lot of mapping, and it takes the
> less time overall to do an
> area from good imagery first, then go fill in the details on the
> ground than to do it all from tracing.  It also makes the ground visit
> quicker, ....
>

As I said, I recognise there is a divergence of views here, including among
people who have made substantial contributions to the map.  Most of the
views have been given a fair airing in the past, and I'm not expecting a new
consensus here and now.  However, I do, with respect, still disagree with
you.

As you say, time isn't the only consideration.  I wouldn't want to be
navigating anywhere important based on a map merely consisting of vectorised
aerial imagery.  IMO OSMers are the ones who should be having the adventures
down the road that may or may not connect, may or may not be open to the
public, etc, leaving our data consumers with the benefit of our endeavours
with maps accurately reflecting what is on the ground.

Ian.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-au/attachments/20110906/7852a27e/attachment.html>


More information about the Talk-au mailing list