After 34 years, the Linux kernel community finally has a contingency plan to replace Linus Torvalds — formal plan drawn up now community is 'getting grey and old'
Eeny, meeny, miny, moe...
Get 3DTested's best news and in-depth reviews, straight to your inbox.
You are now subscribed
Your newsletter sign-up was successful
Linus Torvalds, creator of the Linux kernel, has been its lead maintainer since its inception in 1991. That's a long way back now, and in the man's own words, the kernel community is "getting grey and old." Interestingly enough, though, it was only few days ago that a formal plan was drawn to replace Torvalds, should he wish to retire or something happens to him, or whoever else may be in charge.
The contingency plan now in place is fairly simple, and only triggered if there's not a graceful transition when the occasion arises. Should there be a need for it, the kernel community will first find an Organizer, who is the last Maintainer Summit organizer, or the current Linux Foundation Technical Advisory Board Chair (TAB).
The Organizer has 72 hours to kick off discussions with the invitees of the most recent Maintainers Summit. Should 15 months have elapsed since the last Maintainers Summit, then it's up to the TAB to determine the invitees, who may bring other maintainers in as they see fit. This collective then presents its decision in two weeks' time and communicates to the community via mailing list.
If push came to shove, even without this official plan, it's likely that the kernel community would easily reach an agreement. After all, as Torvalds himself remarked in the past, there are not a lot of open-source projects that "have maintainers that have literally been around for over three decades".
Despite expressing some concern about the number of maintainers earlier in the decade, he's not currently concerned about the technical acumen of the talent pool, either, going as far as saying that "it's not instant, but there are new people who come in, and three years later they are a main developer."
This scenario should be familiar to any software project manager, as it illustrates the key concept of "bus factor," or the number of people that can get hit by a bus before your project is in trouble. Right now, that figure is a nice round zero for the Linux kernel, despite the community being good at self-management. For the time being, Linus Torvalds hasn't expressed any desire to stop being the Linux kernel head, but any increase in the bus factor is welcome, especially coming from nil.
Follow 3DTested on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.
Get 3DTested's best news and in-depth reviews, straight to your inbox.
-
abufrejoval I've always felt and argued that Linus' primary quality is his ability to produce good social code via decision making and supporting a culture, rather than programming computer code.Reply
One of his first such decisions was to recognize that others were better at writing operating system code and to offer his code to full replacement.
And he is now repeating the very same concept on a meta level, just what I'd have expected and I quite sincerely admire.
I don't praise easily: they say it spoils character... -
praz01 Reply
"Recognize that others were better at writing operating system." I'm going to take this sentence purposely out of context to mention that Torvalds' notorious for rejecting pull requests (code changes), on some occasions lashing out in less than civil manner (non-pc)! The Linux Foundation code of conduct solely exists to keep Linus' in check.abufrejoval said:I've always felt and argued that Linus' primary quality is his ability to produce good social code via decision making and supporting a culture, rather than programming computer code.
One of his first such decisions was to recognize that others were better at writing operating system code and to offer his code to full replacement.
And he is now repeating the very same concept on a meta level, just what I'd have expected and I quite sincerely admire.
I don't praise easily: they say it spoils character... -
abufrejoval Reply
I knew this was going to come;-)praz01 said:"recognize that others were better at writing operating system." I'm going to take this sentence purposely out of context to mention that Torvalds' notorious for rejecting pull requests (code changes), on some occasions lashing out in less than civil manner (non-pc)! The Linux Foundation code of conduct solely exists to keep Linus' in check.
But even there he reformed much better than I might have done in his place.. So there! -
bit_user Reply
I've read several of his rants and they generally make sense to me. I'd rather someone be in charge who's not afraid to give the blunt, honest truth and doesn't waste lots of time and words talking around the real issue than someone who's a mere politician.praz01 said:"recognize that others were better at writing operating system." I'm going to take this sentence purposely out of context to mention that Torvalds' notorious for rejecting pull requests (code changes), on some occasions lashing out in less than civil manner (non-pc)! The Linux Foundation code of conduct solely exists to keep Linus' in check.
There's lots of conventional wisdom about projects that are "designed by committee" and it's widely held to be a bad way to go. IMO, if you get someone who's too much of a politician in charge, you end up with something similar.
For all the Linus-bashing, you can't argue with Linux' success. So far, we can only imagine how it would've fared under a different style of leadership. -
DS426 There's Linus bashers and then cult-like followers, with the rest of us somewhere in the middle (or rather not caring at all). Indeed many of us would have much bigger egos and try to benefit personally more from being in this position.Reply
As for this succession planning, I'm surprised this hasn't been done already. Failing to plan is planning to fail, and that could have very well happened. Linus would remain at the top of the pyramid as long as he wishes (and is alive), so there's no real downside to finding and determining successors. Linux deserves and needs the resiliency of an even more robust organization, even as great as it's community structure is.
Plenty of technical competence at upper leadership levels but long-term strategy and vision is still an afterthought (or finally catching up as the "gray old man" and others ponder these things), even as the Linux Foundation's aim was to provide those things. -
bit_user Reply
Yeah, I'd have expected it to be near the top of Linux Foundation's priorities, upon its founding.DS426 said:As for this succession planning, I'm surprised this hasn't been done already. -
randomizer Replypraz01 said:"recognize that others were better at writing operating system." I'm going to take this sentence purposely out of context to mention that Torvalds' notorious for rejecting pull requests (code changes), on some occasions lashing out in less than civil manner (non-pc)! The Linux Foundation code of conduct solely exists to keep Linus' in check.
Others were better, but not those others.:) -
abufrejoval Reply
Last time I clashed with Linus was when I was working on my CS thesis, writing a Unix emulator for a micro-kernel OS that unfortunately wasn't open source.bit_user said:I've read several of his rants and they generally make sense to me. I'd rather someone be in charge who's not afraid to give the blunt, honest truth and doesn't waste lots of time and words talking around the real issue than someone who's a mere politician.
There's lots of conventional wisdom about projects that are "designed by committee" and it's widely held to be a bad way to go. IMO, if you get someone who's too much of a politician in charge, you end up with something similar.
For all the Linus-bashing, you can't argue with Linux' success. So far, we can only imagine how it would've fared under a different style of leadership.
At the time Linux was still almost unusable (as Jochen Liedke, working in the same German research organization as I would explain in more detail later), BSD386 was clearly far ahead of Linux in the Unixoid camp (I myself was running a Unixware System V.4 on my 80486), while micokernels were clearly the way to go in the distributed OS environmnet I was working on: this was the day and age of the Transputer where parallelism was moved right into the programming language via Occam, and there was also QNX, Chorus, Mosix and Minix to challenge the monolithic Multics cast-offs.
I don't recall the exact topic, but it must have been philosophical and I was quite clearly off into a far more "federal" distribution of the "God like" view of the self contained Unix-verse.
What I remember is that we met, battled and left the battleground, without reconciling our differences, but with our mutual respect intact, quite in line with the critical rationalism lemma, that many disputes cannot be resolved by reasoning, as they are based on atomic beliefs.
"Agree to disagree" and then try to do our best in our separate fields of responsibility was the vibe we shared, that's the part we'd probably agree on still today.
Being right doesn't mean you'll have a lot of friends: unfortunately you have to decide if you'll spend your limited intellectual budget into diplomacy or solidifying your science. And since that's been true even for true geniuses like Mozart, how can far lesser minds hope to do better? -
abufrejoval Reply
He's preached his succession gospel for decades, the principles were well established and he didn't deviate. It may just not have been as public and explicit.DS426 said:As for this succession planning, I'm surprised this hasn't been done already. Failing to plan is planning to fail, and that could have very well happened. Linus would remain at the top of the pyramid as long as he wishes (and is alive), so there's no real downside to finding and determining successors. Linux deserves and needs the resiliency of an even more robust organization, even as great as it's community structure is.
All that was done was solidifying those principles into actual near algorithmic rules, since he judged that the probabilities around his ability to continue the job were shifting: again, another sign of his ability for excellent judgment in my view.