On Mon, 06 Oct 2014 21:41:20 +0200 Clemens Ladisch clemens@ladisch.de wrote:
Andreas Mohr wrote:
I think I have an idea which might be useful to accept: for every piece of sufficiently "vintage" submission, people would be tasked with offering (or somehow ensuring) a sufficiently closely time-related cleanup in other places.
The problem that such a new driver imposes is not a one-time reduction in overall kernel quality, but the ongoing maintenance effort.
Vintage is not IMHO a useful test. We have plenty of vintage hardware with active hands-on maintainers which causes no problem, and plenty of modern drivers with basically no maintainer that causes endless problems.
I think there is much merit in the Debian approach - if it's not got a maintainer, and nobody is willing to take it on, throw it out. If it's got a maintainer leave it in.
Economics will resolve the rest of the problem reasonably efficiently. If someone cares enough about using it then they'll figure out how to keep it maintained.
Alan