On Thu, Oct 11, 2012 at 1:31 AM, Maarten Deen <md...@xs4all.nl> wrote: > On 2012-10-10 22:22, Richard Weait wrote: >> >> What are the results?
>> I've taken a go at graphing comment quality. This is not science of >> course. Try it yourself; how would you measure comment quality? > The question is: how do _you_ measure comment quality? These numbers say > nothing to me. You must have some formula to convert a comment to a quality. Yes, I do. I consider comment length and unique comments. There are certainly better ways to measure comment quality. How would you do it? My formula strips the digits from all comments. That penalizes the editor that uses a bounding box as a comment, but I'm fine with that. A bounding box is a terrible comment. :-) Then I count comment characters, unique comments and comments only since API 0.6 was deployed. There were no changeset comments before that. My formula considers comment length and uniqueness. There is probably room to consider comment rate as well. > So high = good quality? Even that was not clear. Yes. What is unclear to me is where the floor should be for good comments? 5? 10? 20? It will vary by situation, obviously. One bad comment doesn't make a bad mapper. Don't take this too seriously. _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
