[talk-au] Re-Mapping (RE: various posts)
richard at weait.com
Sat Jul 21 01:58:33 BST 2012
On Fri, Jul 20, 2012 at 7:51 PM, Paul HAYDON <cadmanager at live.com.au> wrote:
> Hello all,
> A couple of comments regarding recent posts:
> > Date: Fri, 20 Jul 2012 07:21:38 -0400
> > From: Richard Weait <richard at weait.com>
> > To: OSM Australian Talk List <talk-au at openstreetmap.org>
> > Subject: Re: [talk-au] Redaction progress
> > > What's worse - a town that is perfectly mapped but 100m, or
> > > the same town not mapped at all?
> > Well, this is OSM, so there will be well reasoned arguments for both
> > of the alternatives that you present and several that we haven't
> > imagined. :-)
> > Best regards and happy mapping,
> > Richard
> As Richard says, there will be arguments for and against. Just an observation on my part, but be wary - in my experience (in industry) temporary fixes which are put in place can often be left "as-is" later on, despite that never being the intention at the time of implementation. If we map with known errors, how are their updating/correction going to be managed?
This is a discussion that comes up in OSM fairly often. It can be
ably and entertainingly argued over a few beverages. Often the topic
is framed as "even a bad import is better than no data in
specific-area-of-interest". The question of how to update or correct
that data is also largely unsolved. Ideally the answer is, "future
mappers will fix it" but practically that seems too happen less-often
than we would like. An example is the TIGER import in USA, where some
areas have the known-imperfect data, untouched after a few years.
> Perhaps a tag to identify them if known/suspected errors are included, or simply to be verified?
Sure. Tags on the object or on the changeset can help other mappers
or your future self understand the limitations of your earlier
mapping. As an example, I've mapped shopping malls based on rough,
hand sketched outlines, and left a note like "building positions are
estimated". Later trips with a GPS to verify, or after improve
imagery becomes available make it possible to improve the data.
That's a slightly contrived example of the data being provided with
known deficiencies, then improved later. This seems to be the limit
of the scale where this works; within the natural range of a local
Best regards and happy mapping,
More information about the Talk-au