On Thu, May 23, 2013 at 08:28:57PM +0200, WanMil wrote:
Hi Minko,
it would be possible. But I fear that this would be the start point for a bunch of rules which names match and which names do not match:
Main Street MainStreet Main Str. Main St. Main main street Mein Street Str. Main Main Road ...
Sure about the "Mein Street"? :-) This is a bit in the "garbage in, garbage out" category. What about very similar street names, such as Mäyräkuja and Myyräkuja that are very close to each other? Or lots of $basename{tie,kuja,polku,rinne} next to each other? Or, in Estonia, you should not mix the street name suffixes tn and mnt. I wonder if there is any misspelled "nt" that could be either "mnt" (first letter missing) or "tn" (transposed). I am against automatic "correction" of the data. If this is implemented, it should be issuing warnings, and map makers should either fix or blacklist those warnings on a case-by-case basis. (By blacklisting, I mean something like "grep -vf blacklist.txt mkgmap.log.0" and adding suitable regexps to blacklist.txt to suppress bogus warnings). Best regards, Marko