On Wednesday 12 January 2011 15:04:29 Malcolm Herring wrote: > On 12/01/2011 02:36, Steve Singer wrote: > > Have you considered spatial matching? For each light you want to import > > check to see if there is already a light within x meters (say 10-20?) if > > so excluded the light from the import and leave it to be manually > > inspected? > > Given that LoL positions are only specified to within 0.1', a duplicate > search radius would have to be more like 300m to take into account > positional inaccuracies of both. Within such a radius there would often > be more than one other light, particularly in buoyed channels and on > pierheads. > > There is also the difficulty of identifying a duplicate. Matching of > names would likely result in a high error rate. Other features such as > buoy type or colour may also be problematic. Therefore framing and > writing matching criteria rules would be a huge task. > > Given that the duplicate problem would be concentrated in areas that are > actively mapped, it might be easier to identify the mappers and > circulate a notification and request them to re-visit & edit their > previous mapped areas.
Yes, everything is exactly like you wrote. But I will work on that post some more information about duplicates, problems and hopefully solutions ;) Regards, Bernhard
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Imports mailing list [email protected] http://lists.openstreetmap.org/listinfo/imports
