Posted Sep 12, 2008 1:31 UTC (Fri) by kirkengaard (subscriber, #15022)
Parent article: LIRC delurks
Hooray for progress. Here's to more progress, so that what gets merged is worth keeping around. User base does not assure developer time for code fixes when something is functional, and badly-designed/sub-optimal -- but functional -- code has no better life in-tree now than it ever did. Cf. devfs for the canonical example, albeit with what sounds like less active developer interest. Do we have promise of active and ongoing maintainership?
Haven't we been talking about raising the quality of kernel code? Isn't this what -mm is for? Does anyone remember the flack that Reiser4 got for bad interface paradigm and mis-features in the code? Was that all just because Hans was an ungracious, arrogant, abrasive force in the community? Don't we work with developers who are willing to work with us through the external patch trees to get the code to be suitable for the mainline, and then ship it?