[Talk-us] Fresno castradal imports

[Gregory Arenius]

I think that this import was poorly done but overall I don't see the
problem with having this type of data in OSM.  In fact, I think it is
really useful and
don't entirely understand the backlash against it.  I would really enjoy
having that kind of base to work with.

So, Thoughts on Plot and Building data in OSM.

  A.  Great for rendering.  (Generaly this is probably why all these
imports happen the way they do.)
    1. The way Mapnik displays landuse data as different colors allows
renders to display useful and interesting information visually at a very
fine grained way.
    2. Some maps render the plots directly and it can look very nice.  For
instance, Google Maps renders the parts of San Francisco this way with the
plots as very faint lines.  It looks great
      and conveys more information.
         i. Yes, this could be done by rendering OSM and then rendering the
plot data separately.  This could, however, be said of everything in OSM.
The OSM model isn't everything in a separate layer, it is
           everything altogether.  What makes plots unique in this regard?
Why do we want them separate?

  B.  Editing.  (I think the problems here are part of the reason for the
hostility towards imports, especially ones of this type)
    1.  Building and plot information greatly decreases the download area
because of the 50K object limit.  Some people like to do edits over large
areas and this makes that difficult.
      For instance, in this thread someone mentioned they have 68000 square
kilometers they want to work on.  I edit almost exclusively in a city that
is ~120- square kilometers of land.
      I don't think we're going to get a whole lot of consensus on how
dense the data should be with these types of differing use patterns.  I
would look at having all the building outlines
      as a godsend while others look at it as a problem because they can't
download an entire area as easily.

      Personally I take this as an indication that our tool chain really
needs more work.  The 50K object limit is what is getting in the way of
mappers with this problem, not the data itself.
      Even in San Francisco where, other than the streets and some GNIS
nodes, there really hasn't been any imports this limit gets in the way.
How do mappers in places with really dense non-imported
      data deal with this issue?

    2.  Dense building and plot information greatly increases the file size
that editors have to cope with and many of them can't, especially for large
areas.  Again, I don't think a consensus
      is going to be easy to find here.  I want dense data, building
outlines and plots.  I think that is useful for many purposes.  Others
aren't interested in this type of data and find it nothing
      but a nuisance.  Again, maybe better tools will help but I think
we're more likely to see this type of data multiply rather than disappear.

    3.  This one is more subjective but the first thing I thought when I
zoomed in on the Fresno data was "awesome."  I get the impression other
people (most?) think "eww, what a mess!"
      I would enjoy that kind of base to work with and expand on even if
imperfect.  Building outlines would be better but plot information would be
great, too.  Mostly it would just be so much
      easier to tag a pre-existing area instead of having to trace one from
scratch every single time you want to add a park, or school, or fire
station etc to the map.  If the area didn't look right,
      edit it.  I think people look too much into the plot data being
"official" and "owned by someone else."  Once its in OSM its ours and we
own it.

This import does have its problems but whats the rush to delete it?
Without imported data Fresno would pretty much just be a white place on the
map.  The streets are all tiger import, many of the POIs are
from the GNIS import and almost all of the rest of the data is from this
one and little of any of it has seen much editing.  Are we really making a
more useful set of geodata by cleaning out
the plot info here?

Cheers,
Greg
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20120504/b5ff8837/attachment.html>