In case anyone is working on a large-scale TIGER09 import, here are some things I'm noticing in the data:
A. Spanish street names are still not always proper-cased. IME, it is generally the practice in the US that street names are written as "El Morado", not "el Morado"[1]. B. Some street names are being abbreviated in ways that the average consumer may not recognize. Examples[1]: Cll Quebrada (Calle Quebrada) Cam Norte (Camino Norte) Also, suffixes are sometimes abbreviated strangely, like "Magdalena Ci" instead of "Magdalena Cir" (Circle). At first glance in a small font, I mis-read it as "Ct"[2]. C. In some places, it seems like they may have gone overboard in their newfound quest for accuracy, creating far more points than necessary to accurately represent a curve at a scale likely to be used by a consumer. One example I'm looking at[3] uses 23 points for a 90-degree "knee-shaped" corner with a total arc length of about 100'. They aren't even following the "knee" - it's drawn with a constant radius (about 55'). I'd have used between 1 and 5 points[4] to model the same turn. Depending on how prevalent this is elsewhere, my concern is a significant increase in the number of nodes in the database. It may be good to "simplify" such ways during import. [1] Examples from San Bernardino County, CA (tl_2009_06071_edges.shp) [2] Example from Riverside County, CA (tl_2009_06065_edges.shp) [3] Peony St/Hyacinth Wy in Riverside County, CA (tl_2009_06065_edges.shp) [4] Depending on how busy/creative I feel :) -- Alan Mintz <[email protected]> _______________________________________________ Imports mailing list [email protected] http://lists.openstreetmap.org/listinfo/imports
