Re: [OSM-dev] Chinese spam diaries, an analysis
On 03.12.14 17:14, Andy Allan wrote: Thanks for the analysis, I hope it provides developers with ideas for combatting it via the automated spam filters that we already have[1]. I'd suggest to extend/refine the automated filter somewhat. Say: * a novice ist not allowed to post at all * a novice who did some changesets is allowed to post say once per day * an intermediate is allowed to post say once per hour * for an expert (either subscribed for years or lots of changesets) the posting limit is waived One could even think to allow experts to delete other user's posts (because of spam). Of course a log has to be maintained. And so no special people (moderators etc.) are needed! And of course the parameters need to be optimized: * how long is a user a novice? * is 10 changesets enough to allow him/her to post? * when does the intermediate level start? 2 years? 100 changesets? * what are the achievements to reach the expert level? 4 years? 1000 changesets? Those parameters could be tweaked on the fly, I'd say. /al ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Chinese spam diaries, an analysis
On 04/12/14 11:17, Andreas Labres wrote: On 03.12.14 17:14, Andy Allan wrote: Thanks for the analysis, I hope it provides developers with ideas for combatting it via the automated spam filters that we already have[1]. I'd suggest to extend/refine the automated filter somewhat. Say: * a novice ist not allowed to post at all * a novice who did some changesets is allowed to post say once per day * an intermediate is allowed to post say once per hour * for an expert (either subscribed for years or lots of changesets) the posting limit is waived So in other words, most of things we already factor in to our spam scoring... We're just not quite as rigid. In particular you can still post (within reason) without having made any edits - it is actually surprisingly common for non-spammers to do that. Tom -- Tom Hughes (t...@compton.nu) http://compton.nu/ ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Chinese spam diaries, an analysis
On 03/12/14 16:14, Andy Allan wrote: However, spam is an arms race, and I think we might need a different long-term approach. I know in the past using 3rd-party spam filtering services was too expensive (and not really very OSM-ish either). The main such system is akmiset and I'd love to use it but (a) it costs money and (b) to make it most effective we would have to send it things like email addresses and IP addresses which I figure people may object to. Perhaps we need a new set of human content moderators on the site, say 40-80 people with a variety of languages between them. We can consider grey-listing all accounts - i.e. the first few posts of every account is held for review automatically by default, and enable direct posting after we're more certain they aren't a spammer. Once we have a review queue and moderator system then obviously it becomes trivial to do things like holding posts from new users for moderation - we need the basic infrastructure first though. Tom -- Tom Hughes (t...@compton.nu) http://compton.nu/ ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Chinese spam diaries, an analysis
On 04.12.14 12:33, Tom Hughes wrote: So in other words, most of things we already factor in to our spam scoring... We're just not quite as rigid. A (hidden) spam score is bad (IMO). Nobody sees it, almost nobody can test it. A documented user level with documented rules would make much more sense and (IMO) would much more likely be accepted. In particular you can still post (within reason) without having made any edits - it is actually surprisingly common for non-spammers to do that. OSM is not a blog site. OSM is about making the data better. Once you have somehow figured out a little bit how OSM works, you could blog about it. IMO. /al ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Chinese spam diaries, an analysis
On 04/12/14 12:06, Andreas Labres wrote: On 04.12.14 12:33, Tom Hughes wrote: So in other words, most of things we already factor in to our spam scoring... We're just not quite as rigid. A (hidden) spam score is bad (IMO). Nobody sees it, almost nobody can test it. Nothing is hidden: https://github.com/openstreetmap/openstreetmap-website/blob/master/app/models/user.rb#L210 Tom -- Tom Hughes (t...@compton.nu) http://compton.nu/ ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev