Graham Leggett wrote: > On Wed, November 29, 2006 7:46 am, Stefano Mazzocchi wrote: > >> Yes, but a lot of effort needs to be put in place to avoid creating >> 'altars' that people want to create races about. >> >> For example: most active developer, most active project, most email->svn >> activity... all these will turn out to be sources of flamewars or social >> friction. >> >> We can surely apply these 'metrics' to the various projects and show >> 'quantitative' results and be creative on which things to measure, but >> to suggest or even hint that these might be "really" a measure of >> quality, I would be very strongly against that as it might create more >> harm than good. > > I have seen this topic come up before a few years ago, as I recall it was > stats on the [EMAIL PROTECTED] list. The general consensus then was that stats > like this were generally a bad idea, for the same reasons you list above.
Right, the status quo in the ASF is that 'metrics on community health are a bad idea' because they generate races and they tend to polarize or flatten a very faceted and multidimensional aspect into just a few, potentially harming the effort. *but* we are labs and we have all sort of disclaimers now about all this being tries to improve the situation. For example, Agora is a community analysis tool and I've used it successfully to stop fights between people by showing them how their traction in the community was (go to http://people.apache.org/~stefano/agora/, load the mail lists that you care for from the side bar, load the graph clicking the bottom left button, start the graph, then double-click on any two nodes and drag them apart, you'll see which one has the most influence on the community) I spent endless hours thinking on how to show community shape and status without showing any number at all: it would have been easy to 'rate' people based on their 'distance from the community centroid' or even fascinating to see how such a distance would change every month, the runner ups and people being pushed away from the center. My fear was that even if the agora algorithm, being based on topology analysis, is more robust to attacks or gaming than, say, number of email per month, it is still pretty trivial to game. And I know that computer geeks have a natural tendency to want to climb figures, when they are presented, especially if very publicly, because they might harm their egos. Agora has been around for years and *never* created any problem in the 'ego race' respect, and I've had several incubator mentors complaining when it was down and not updating the data. So, in this respect, I welcome MBoxer or any other tool in the direction of exposing more information out of mail list or general project activity... *but*... the impact on the social evolution of the project must be a design constraint up front, not just a polite afterthought. -- Stefano. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
