[Imports] Value of imports in general

[Ilpo Järvinen]

On Fri, 7 Feb 2014, David Marín Carreño wrote:

> Hi all.
> 
> 2014-02-07 12:29 GMT+01:00 Johan C <osmned at gmail.com>:
> > Imports are bad when: "In this case, an import cannot happen when there is
> > no local community to keep it actual."
> 
> I completely disagree.
> 
> When there's no local community to keep actual the data in the
> database, an import is as bad as any manual edition made by an
> ocassional collaborator.
> In both cases, we will have OSM reflecting a maybe-outdated state of
> the reality.
> 
> The problem here wouldn't be if the data come from an import or from a
> sporadical collaborator. The problem is that there are many places in
> the world in which there's no local community. And, IMHO, it would be
> better an outdated map than having no map at all...
> 
> And, for a sporadical collaborator it would be always easier
> correcting an outdated element in a map (as long as the imported
> data-structure is good) than creating it from blank.

I totally agree. Lots of mapper time is wasted on drawing geometries 
or fixing details that were derived from poor GPS traces (ironically I 
could give some superlatives here about how precious some consider that
this particular garbage ;-)). Most of those geometries could just as well 
have been imported and have equal or even better quality, assuming that 
the source quality is minimally reasonable and the import is properly 
executed.

This comes with the experience of way beyond 10k+ addresses, 10k+ 
entrances and lots of other details surveyed (now including even most
of the traffic signs since I've long since ran out of more useful things 
to map within the usual area I map). It also implies that details close
to me are pretty complete but always when I end up visiting a new area
I encounter this very same issue that I need to fix a lot that usually
comes from previous mappers drawing things wrong or having their device 
mislead them!


-- 
 i.