Are all patches created equal?
Are all patches created equal?
Posted Jul 21, 2011 18:58 UTC (Thu) by david.a.wheeler (subscriber, #72896)Parent article: Are all patches created equal?
If you can make better metrics, that'd be nice, but even these "unfair" metrics still help give some insight. The fact that Microsoft was working to get its contributions into the *mainline* Linux kernel was far more obvious from the statistics. So even though all measurements have weaknesses, they're still useful, and I'm glad LWN is doing them.
Posted Jul 21, 2011 20:38 UTC (Thu)
by jspaleta (subscriber, #50639)
[Link] (1 responses)
As a community we value devs you "do the right thing" and mainline their new ideas. Can we capture that as a broken out metric? Activity in staging broken out would be the thing to look at for that.
As a community we strongly dislike un-maintained code dumps. So how can we capture that and produce either a stat that lifts up good maintainers or a shaming stat for bad maintainers of existing code? Do we trend commits that attempt to fix reported bugs for this metric?
What other specific behaviors do we really like to see as best practices?
Also perhaps its time to move away from fine scale ordering and instead think about chunking stats at a less granular level. No more simple rank ordering, but group highly productive devs/companies, into pools which equate to a curved grade scale weighted to a median value. X number of companies/devs get A grades (doing great) in the maintainership stat, Y companies get C grades (doing poorly), Z companies get B grades (so-so) and just report that. Stop with the micro-managed class ranking and have a pool of honors students that can be viewed collectively as peers who set the standards for best practices.
-jef
Posted Jul 21, 2011 21:49 UTC (Thu)
by dlang (guest, #313)
[Link]
how to the contributions bunch.
we've been looking at the top few in detail, but how does it taper off?
Are all patches created equal?
Are all patches created equal?