On Mon, May 2, 2011 at 11:47 AM, Nathan Edgars II <nerou...@gmail.com>wrote:
> On 5/2/2011 4:29 AM, Martijn van Exel wrote: > >> On Mon, May 2, 2011 at 9:22 AM, Nathan Edgars II <nerou...@gmail.com >> http://wiki.openstreetmap.org/wiki/File:West_Harrisburg.jpg >> As such, it could help >> reduce - though never fully prevent - cases like the one you link to. >> > > How is that? West Harrisburg was created using ordinary editing tools over > a matter of months. Certainly you're not proposing that new editors can't > draw roads? That could be an option, although you'd need to be able to move on pretty quickly from there. Reputation systems are about trust: you gain trust by delivering quality contributions, you lose it by delivering crap. On eBay it involves peer-rating of contributions, on StackOverflow it's a combination of peer-reviewing and badges, and on OSM I think it could look something like that. You would get badges for doing good things, and changesets could be peer-reviewed (though probably only by fellow users who have done any significant amount of editing in the same area - or at least their ratings would weigh in more). On StackOverflow you can do useful things right from the start: ask and answer questions. Once you gain rep, you can start doing even more useful things quite soon: vote contributions up or down, edit questions, create new tags. I think the benefits of higher quality data and less churn outweigh the imposed limitations associated with the learning curve. > > >> What is the story behind this West Harrisburg vandalism situation by the >> way? >> > > "Didn't realize this map was supposed to be a real map. Just through[sic] > some stuff in there. Thought it was a game or something. Come to find out > it's more like Wikipedia." > > Creating "badges" for a certain number of edits or other such grinding > would, if anything, increase the amount of bogus editing as new editors try > to get those badges, and would piss off new editors who know what they're > doing (for example, soon after joining, I imported the complete NYC Subway > system from a shapefile I had previously created by tracing USGS topos, and > successfully merged the portions that had already been mapped). > Probably one of the main challenges for a rep system is getting the balance just right: you want to stimulate doing good things but reduce collateral damage (badge hunting) as much as possible. You wouldn't want to reward pure quantity - anyone can add 1000 nodes in one day. Does that mean it can't work at all? I don't think so. In this particular case the threshold could be a little more complicated with a temporal dimension - the Node Regular badge for 1000 nodes in a month, with no more than 100 on any given day. I mean, there's ways to tackle these problems - as there will always be ways to beat the system. You're not going to eliminate evil. As for your specific case of importing the NYC subway - that's great work, and you apparently knew what you were doing. A lot of people don't, and we see the negative effects of that more and more often, pissing off (and even scaring away) experienced mappers. Who would you rather lose: committed and experienced mappers frustrated with yet another ill-advised import, or those not willing to take the time to address the learning curve that is inherently there in an open, do-ocracy-based community? -- Martijn van Exel http://about.me/mvexel
_______________________________________________ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
