LCA2009: That mysterious thing called the kernel

Sam Varghese
Wednesday, 21 January 2009 14:23

Opinion and Analysis

For the last four years, the pony-tailed Jonathan Corbet, kernel developer and editor, has presented what he calls the kernel report at Australia's national Linux conference.
This year, he did not do so - but he had something equally interesting to offer at a session in Hobart today, the first day of the 10th conference proper. (Monday and Tuesday were devoted to the mini-conferences).

Corbet's talk, titled "Joining the mob: a guide to the kernel development process", was presented in his usual concise manner.

He is a man who displays an admirable economy with words and the language he uses is simple and to the point. Much like the kernel creator himself, Linus Benedict Torvalds, who was among the audience, occasionally contributing answers to the questions that the audience raised.

Corbet began by outlining reasons why participation in the kernel development process matters - the kernel, he said, was the core of the Linux system and therefore should continue to be the best it could.

Participation was the way by which people could get the kernel to meet their needs; complaining from outside would probably get something of a hearing but those on the inside were more capable of getting their needs met as they could often end up making the very changes that they needed, he said.

Corbet pointed out that maintaining code outside the mainline kernel was expensive; if code made it to this kernel, the one that is handled by Torvalds, then the chances of it being in every distribution were very good.

CONTINUED