On Friday 05 Apr 2013 13:56:11 Steve Dougherty wrote: > > Would it make sense to have nodes insert bundles of metrics on how > they used the setup? That could be really helpful, as long as it could > be done without revealing identifiable information. I mean things like > how long did they take overall, what options did they switch between > before deciding, what settings did they choose? It should of course be > opt-in, but I'd prefer for statistical reasons that someone who didn't > opt in inserts a non-response marker so that we can have an idea of > response rate. We could guess based on download vs. opt-in rate, but > that's potentially much less precise. > > Irfan: Do you have suggestions on collecting interface usage metrics?
Interesting idea. The main problem with it is it wouldn't include the users who didn't complete the setup wizard. Plus spam is a possibility, as with all these things... We could limit that by making it a probe request, but then it's not anonymous enough? Depends on what exactly you are collecting I guess ... We could collect data including failures, but only if we ask the user at the beginning, which would give a bad impression... We could only ask a randomly chosen subset of users?
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Devl mailing list [email protected] https://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
