kernel upgrades and kernel driversmodules dependancies question 4807

This is inaccurate, and notoriously so: They decline to keep the driver interfaces fixed because that would result in a variety of serious technical suckage, that would make life for them unpleasant and impede the advance of Linux. Details here:

Have you ever considered why they change? Did it dawn on you that it was for technical reasons, not "political"? No, I guess it really didn't.

This is also inaccurate. The driver-ide drivers continue to treat mbutt storage devices attached to supported SATA chipsets asdev-hdX devices. It's only Jeff Garzik's additional libata driver colleciton that addresses such devices asdev-sdX.

Again, have you ever bothered to inquire as to why that is the case -- before going around rather ignorantly declaiming that "kernel interfaces mutate", which they did not do at all in this case, let alone "rapidly"?

It's because Garzik was able to leverage the existence of the kernel's well-tested SCSI layers for his driver set, which made those SATA drivers easier to write and better-performing than the older drivers-ide ones.

Now, that is actually someone relevant to your intention. sys got created becauseproc has become a hopeless mess, and starting a new tree was easier than cleaning up the old one.