[HOT] Central African Republic import of WRI data

Severin MENARD severin.menard at gmail.com
Thu Jul 4 03:05:37 UTC 2013


Hi,

I agree with Andrew regarding the importance of creating a connected road
network. Actually I already mapped a lot the roads with Bing Imagery and
some sections are not covered by high res imagery but sometimes the LandSat
medium res allows to identify the roads. In these cases I traced it and
informed about the bad accuracy in a fixme.

Sincerely,

Severin


On Tue, Jul 2, 2013 at 12:28 PM, Andrew Buck <andrew.r.buck at gmail.com>wrote:

> I understand the 'bad data worse than no data argument' but I think that
> only goes so far.  Having a road on the map that isn't terribly accurate is
> unlikely to cause major problems for people.  Having a hospital where none
> exists could be a major problem, but for roads I think this is less of an
> issue.
>
> Either way of doing the import is fine by me, but my preference would be
> to get the data in (especially if there are names or other metadata) and
> then just refine the location later on.  We tend, as a community, to focus
> way too much on positional accuracy while we omit focus on other forms of
> accuracy (topological, metadata, etc).  If the only thing wrong with the
> WRI data is the positional accuracy I don't see a problem.  Especially as
> it is being imported into a largely empty area of the map.  200 meters
> error in the jungle in empty area is no big deal, 200 meters in a densely
> mapped urban area would be a much bigger problem.
>
> Anyway, just my 2 cents.  I will defer to the community's assessment of
> how to do the import, especially since someone else is doing the work.  ;)
> But I thought I would at least share this insight.
>
>  -AndrewBuck
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/hot/attachments/20130703/71a305eb/attachment-0001.html>


More information about the HOT mailing list