On Mon, Apr 25, 2011 at 2:37 PM, Frederik Ramm <[email protected]> wrote:
>> Why not programmatically do the merge? Why must it be manual? > > Let's put it this way. If you spend a month doing this manually for the > whole country, then it's not better. But by preferring manual operation, we > at the same time usually prefer local operation - a local mapper is likely > to pick up his "parcel" of data and manually work on that. That mapper will > then have a sense of ownership (maybe, thinking of the recent license change > discussion, a concept like "parenthood" would be more appropriate) towards > the data, and the mapper is more likely to care for the data in the future > than some hacker from another continent who thought he had a cool script. My preference for data entering OSM is the following: 1. Personally surveyed 2. Traced from up to date imagery 3. Imported from a highly reliable source Why imagery over imports? Because we don't currently have a good way to process all the data in photographs in a way that makes sense. Maybe one day we'll have the kind of technology that these imagery companies brag about, where we'll be able to analyse a photograph and pick out every features of the road, look at the direction of cats, read the signs to determine road names, etc. But right now we can't even do simple data merges. >> I'm just arguing against the constant maligning of imports. I want to do >> an import in the future, but I also want to be good citizen and be familiar >> with the approaches and results. What I'm observing more often than not is >> a general perception that imports shouldn't be part of OSM. > > Yes, because imports generally give us data without a soul, whereas a mapper > gives us data with a soul. It's scary when a German gets touchy-feely, isn't it. Let's be more practical. One might argue that like Wikipedia, there will be a point at which OSM is entirely complete and the project will go into maintenance mode. We're a long ways away from that right now, but it's possible it might happen. If it were, what we'd want would be a way to keep the existing data in OSM up to date- not just administrative data, but data that doesn't appear anywhere else, and features which may exist but may not be in the free administrative data yet. For that, we need community. This is especially important in a place like the US, where our mapper density is low when compared to our European friends. The question which comes back to the project every so often is this push/pull between using existing datasets, and not. Some argue that by not having the data, by having a blatantly empty map, we encourage more mappers, and that these mappers know that the data is from other community members. Others argue that if the data is available and usable, we should use it. They might also argue that by having /some/ data in the database, it encourages people to map more. Here's the reality: There's no way to really know at this point, at least as far as the US is concerned. Anyway, back to the point of the mail... I agree with Fredrik, at least in that I think humans should be the central driving force in OSM. I'm not against all imports, but nearly all. - Serge _______________________________________________ Imports mailing list [email protected] http://lists.openstreetmap.org/listinfo/imports
