The state of California has some good landcover shapefiles on the Department of Forestry and Fire Protection site. They are sorted by county. The smallest are under a meg while the largest - Fresno - is more than 400 megs. The average size is around 30 to 40 megs. These would be a great addition to OSM since they contain valuable metadata. Once the areas have been added, the state will take on a similar look to states like Georgia and Massachusetts that have had statewide imports done. Another good example of what California can be is the Corine Land Cover (WikiProject Corine Land Cover).
There are several challenges with the data. Here are a couple. * It is a huge dataset and will need some optimization. The straight lines have points every five meters, creating jagged edges that aren't visually attractive. Josm has a plugin but that's probably not the best way. A spline interpolation would really improve it dramatically. It will increase data size, but it's worth it. Maybe there is a batch mode program available to do that. Mapshaper has a tool to optimize a shapefile by reducing the details in the file. It may not be working though. Here is a rough idea of what the areas look like without being simplified. * Only vegetation data should be used. The urban/residential/water/unknown should be skipped since the quality isn't good enough for this purpose and would just create tons of conflicts with existing data. Also there is/will be better data available. * We can use some filters to split the shapefiles into different features. This is also great to prepare the OSM files for each type instead having it mixed all together. * All these polygons have overlapping ways. This should be avoided because it creates tones of duplicate nodes/ways. Each polygon should be split in single ways and the area defined with a relation. According to the description, mapshaper will do that to. There may be a way to do this with postgis or through a Perl script. Validator will just merge duplicate nodes, but can't fix duplicate lines. This is really a tricky thing that many other updates lack. Efficiency should be considered here. Also editing is easier then. This data shouldn't need much editing after import. Please feel free to add your comments and suggestions on the California Land Cover wiki page. This is also being posted to the talk-us list.
_______________________________________________ Imports mailing list Imports@openstreetmap.org http://lists.openstreetmap.org/listinfo/imports