Since the recent discussion in September and October, this is what I’ve been doing when reviewing roads from TIGER:
When working on a highway that still has the tiger:reviewed=no tag, once I consider the road to be reviewed to my satisfaction, I remove that tag (as I always have), but then if I’m feeling like I can take the extra time to do it, I go ahead and remove all the other tiger tags as well, except I leave the tiger:cfcc tag in case someone wants to refer to it in the future. And I convert the tiger:zip_left and/or tiger:zip_right tags to postal_code.
I feel some more discussion of the zip code / postal code tag is warranted. I’ve been thinking about it since a couple of months ago when I suggested that the information should be kept. And I’m not sure that it really should.
In the United States (as far as I am aware), zip codes are not assigned exactly according to street boundaries, but rather the post office, county land records office, and property developer somehow work together to choose a zip code for each address as that property is developed. The “de facto” situation is that zip code boundaries tend to be at major streets within a city, but the streets are not “de jure” the zip code boundaries. A major shopping complex or manufacturing company may have a different zip code than the surrounding buildings for any reason: perhaps historic if it was the first development in that area and later buildings were assigned a different zip code, perhaps for convenience of mail delivery, or maybe even for the prestige of being listed with a big city address. If it’s a big enough corporation or a university, it may have it’s own zip code, different from anything else in it’s city or town.
Therefore, the tiger_zip tags are actually an oversimplification of reality. The majority of the time, all the houses within a neighborhood will have the same zip code (main 5 digit part) and along any single block of any street all the houses on one side will have the same 4-digit extension to the zip code. However, there are exceptions to both of these rules. I think the most useful reason for keeping the postal_code information on a street would be if the buildings aren’t mapped yet or don’t have addr:postcode tagged. Then when adding that information, one can simply refer to the nearby streets to see what the zip code should be for a building. But there’s no guarantee that’s really the correct zip code.
Here are some useful links that I use all the time:
United States Zip Code Boundary Map shows approximate boundaries between zip codes in the United States.
USPS Zip Code Lookup allows you to look up a zip code for any address. It’s also very useful for verifying exact street names, which almost always match street signs on the ground (but the street signs are more likely to have errors in prefixes or suffixes).
Right after my first post in this thread on Sept. 19th, I started editing a neighborhood right on the boundary where three different zip codes are present. I started changing tiger:zip_left and tiger:zip_right to postal_code:left and postal_code:right on the streets that are presumably the boundaries. But I never finished. It was quite a bit of work to keep it all straight in an area where three different zip codes come together, and who is to say if every building near there really conforms to the presumed boundaries? You would have to survey addresses on the ground and then look up each one in the post office database to see what the zip code really is. I gave up on that. If we do keep tiger:zip information in the highway tags, I think it’s most useful to keep it only in areas that are not right on a boundary.
