I hate to step into this flamefest, but:

  Having traveled around the US, I've been really glad the tiger data is
  there.  Often it seems like there have not been a lot of edits, and
  it's way better than nothing.

  I heard about OSM long ago, and I think noticed the map was blank in
  mass, and found it too daunting.  I had gathered trail tracks and was
  going to make my own maps starting around 1999 (with differential
  even), but never got around to it.

  I re-found-out about OSM last winter, and have been editing when I
  have time - my town is now in pretty good shape, including trails in
  conservations lands.

So all in all I'm very appreciative of the imports having happened
(thanks Dave, Chris, and I'm sure others.)


Regardless, TIGER data has been imported, and the only question is the
way forwward.  The thing that seems missing is an organized approach for
replacing mass imported data with better mass imported data.  I view
this as similar to a revision control system.

Logically, the tiger/massgis/canvec/whatever data lives on a 'vendor
branch', like in CVS.  (I realize the osm db has no such concept.)  When
there is new tiger data, one can say:

  For every way in the old import, if it hasn't been touched, then
  replace it with the way in the new import.

The hard part is defining "hasn't been touched" usefully - someone
adding a side road could perhaps still be merged automatically, with the
side road contacting the replacement way at about the same place.  This
is the equivalent of when cvs can automerge changes to a file in
different lines and decide that this is ok.  We need this for vector
data...  (I know, ENOCODE...)

It might also be possible to get local consensus.  People are naturally
and rightly highly reluctant to damage human-done work.  But if you can
gather the people, look at what might be, and say "yes, on balance
that's way better, let's do it and then pick up the pieces", it's
reasonable to go ahead - even if it would be rude without talking to
people.  Of course, that assumes that people who mapped will answer
messages and discuss briefly, but that seems only fair.

Attachment: pgp6N73SyJ2E9.pgp
Description: PGP signature

_______________________________________________
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Reply via email to