Linux Kernel Development Keeps On Picking UpLinux Kernel Development Keeps On Picking Up

The number of individual contributors to the kernel, and the amount of work being put into it, continues to increase. This is measured with the usual metrics: lines of code, number of commits, number of individual contributors and code-approvers.

Ditto the number of companies making kernel contributions. Said companies may often be bitter business rivals, but they all seem to think the Linux kernel is a valuable collaborative project, and they get a lot out of it for what they put into it.

To make it run well with their hardware (IBM, HP, Fujitsu). Server vendors move that many more units when they're compatible with Linux; this is simple fact.

To make it something with which they can compete for customers (Red Hat, Novell). My feeling is those who turn Linux into an end product have to tread lightly with how much they can commit upstream, lest they give away all the goodies -- but most of the really innovative work in any particular distribution takes place outside the kernel anyway.

To use it as a raw-material substrate for devices (Sony, Samsung, Nokia). This should be self-evident by now, given how Linux's biggest current growth is in the form of such things. The more you can commit upstream and make part of the stock kernel, the less you need to add back in by hand each time you re-roll your own Linux-derived product.