--- El vie, 20/8/10, emijrp <emijrp(a)gmail.com> escribió: De: emijrp <emijrp(a)gmail.com> Asunto: Re: [Wiki-research-l] Polish and Russian vandalism revert tags Para: "Research into Wikimedia content and communities" <wiki-research-l(a)lists.wikimedia.org> Fecha: viernes, 20 de agosto, 2010 18:22 Hi; 2010/8/20 Felipe Ortega <glimmer_phoenix(a)yahoo.es> That's also why we are not currently using the MD5 hash approach (we do need to differentiate among different types of reverts, not only detect them). But you can use MD5 to identify *all* the reverts, and then classify them with their comments. Sure, and I think MD5 is a pretty efficient way of identify reverts. The point is that, once we start looking at comments, it's better to go ahead with them, since in theory we don't need to check against MD5 if we missed some revert. However, we will have to compare MD5 and comments procedures. Comments are considered a fairly conservative approach, though a good proxy, and that's why we selected them. F. Regards, emijrp -----Adjunto en línea a continuación----- _______________________________________________ Wiki-research-l mailing list Wiki-research-l(a)lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

--- El sáb, 21/8/10, emijrp <emijrp(a)gmail.com> escribió: De: emijrp <emijrp(a)gmail.com> Asunto: Re: [Wiki-research-l] Polish and Russian vandalism revert tags Para: "Research into Wikimedia content and communities" <wiki-research-l(a)lists.wikimedia.org> Fecha: sábado, 21 de agosto, 2010 12:07 Also, you can ask in the village pumps in PL and RU. You will get replies from the recent changes patrollers. Will do, thanks for the tip! :-) F. 2010/8/20 Felipe Ortega <glimmer_phoenix(a)yahoo.es> --- El vie, 20/8/10, emijrp <emijrp(a)gmail.com> escribió: De: emijrp <emijrp(a)gmail.com> Asunto: Re: [Wiki-research-l] Polish and Russian vandalism revert tags Para: "Research into Wikimedia content and communities" <wiki-research-l(a)lists.wikimedia.org> Fecha: viernes, 20 de agosto, 2010 18:22 Hi; 2010/8/20 Felipe Ortega <glimmer_phoenix(a)yahoo.es> That's also why we are not currently using the MD5 hash approach (we do need to differentiate among different types of reverts, not only detect them). But you can use MD5 to identify *all* the reverts, and then classify them with their comments. Sure, and I think MD5 is a pretty efficient way of identify reverts. The point is that, once we start looking at comments, it's better to go ahead with them, since in theory we don't need to check against MD5 if we missed some revert. However, we will have to compare MD5 and comments procedures. Comments are considered a fairly conservative approach, though a good proxy, and that's why we selected them. F. Regards, emijrp -----Adjunto en línea a continuación----- _______________________________________________ Wiki-research-l mailing list Wiki-research-l(a)lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l _______________________________________________ Wiki-research-l mailing list Wiki-research-l(a)lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l -----Adjunto en línea a continuación----- _______________________________________________ Wiki-research-l mailing list Wiki-research-l(a)lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l