33 comments

[ 3.0 ms ] story [ 276 ms ] thread
First he should write a book on the architecture of the 5.x and 6.x kernels because I can't find one anywhere.
Is it a common impression that consulting the kernel source has gotten more difficult since 5.x for getting the same knowledge?
He hit a bus before probably. Are we talking about internal or universal serial?
Hah, I recall very well when this was posted. It was my first introduction to the bus factor[1], and I still refer to it from time to time.

Segfault.org had a lot of fun jokes, just look at the front page from that time[2] which includes some future-predicting ones like "Netscape 6 Special Edition To Feature Extra Ads"[3] (not Mozilla as such but).

[1]: https://en.wikipedia.org/wiki/Bus_factor

[2]: https://web.archive.org/web/20010616172138/http://segfault.o...

[3]: https://web.archive.org/web/20010619114515/http://segfault.o...

I was recently introduced to the “lottery factor” as a less morbid alternative. As in, if one of us won the lottery tomorrow and quit on the spot what would happen.
I can kinda see the point there, but in the lottery example there's at least a possibility of making contact at a later date.

The bus is irreversible and sudden.

I'd say the lottery example works for example with skills/knowledge. ("Do we have someone else who can do what that person was doing?")

While the Bus factor is very relevant for issues like infrastructure and access ("The only person who knew the important passphrase or credentials is now lying dead on the pavement.)

The difference between the lottery factor and your bus factor makes a decent proxy for how much people like working there.

If someone has a lottery factor equal total their bus factor, then you expect them to say “screw you” when asked for help, and you should probably work on that.

Didn't Ballmer do a cost-benefit study on this in '07 only to determine that if you struck Linus down then he would become more powerful than anyone could imagine?
If true, would that still hold now? There is so much politics and polarization and agendas - I'm not aware in Linux specifically but in tech and open source generally. Without benevolent dictators I think a lot of projects would fragment very badly and leave big openings for companies like Microsoft. I think we're going to see that as a failure mode in big open source projects more and more.
(comment deleted)
It was a joke.
I think it's a bot. Gross.
Oh yes, I remember that white paper. The conclusion was that if Linus became a Force ghost, he would be able to code 24/7, which would affect the market share of Windows Server 2008.
I'm not really part of the Linux community so I feel a bit confused. It definitely seems like it's meant in jest, though at the time of writing, there probably would have been a real concern around business continuity.

Am I just wildly missing the joke?

(comment deleted)
Yeah it's just a joke, lightly poking fun at contemporary commentary worrying about the viability of open source projects perceived as too centralized around one key person. I don't know how serious the concerns ever were, though keep in mind that the idea of relying on open source software was still novel at the time.

As I recall, a lot of the world wide web seemed to be jestful writing of a similar tone (irreverently humorous but ultimately playfully cynical at most, not darkly cynical as today) in those days.

The data is old. In the account of increased population, traffic and pollution, the study should be reconducted just to be safe. The muffin factor probably has changed a lot since then and even we should propose to include his new favorites.
One from the same timeframe:

What if Chuck Norris gets hit by a bus? Time to buy a new bus.

In the history of software development, has a project ever been doomed by a bus killing the key engineer?
Yes, this is actually quite common. As often as tech companies fail due to poor marketing or poor business, they also fail due to technical debt induced by laying off core developers.
I was thinking of a bus - or any other quick and unexpected fatality - specifically.
I'd probably look at plane accidents.
The crash & loss of executives from Chevron & Pacific Bell on PSA Flight 1771 is such an example.
Layoffs are often quick and unexpected, and the impacted employees are essentially "dead to the company".
I think that ReiserFS was doomed because of the key engineer actually killing.
Not a bus, but TropeTrainer: https://www.inverse.com/input/features/tropetrainer-thomas-b..., to the point where synagogues are passing around computers running old copies of macOS so kids can study fir their bar and bat mitzahs
An interesting story. Do you think he subconsciously wanted the program to die with him, to ensure he would be missed?
That was a super interesting yet melancholy read. Thanks for sharing it
Why the web archive link? (why the post at all OP?)

The original is archived over here on the author's website along with the rest of his segfault writings:

https://www.crummy.com/writing/segfault.org/Bus.html

Because what happens if the original author is hit by a bus?
> (why the post at all OP?)

Who the heck even are you to tell me what tech related posts I can or cannot post here. The audacity lmao.

> Why the web archive link?

Because the web archive post is what I was reading and didn't know the author had their own website called crummy. Thanks for the link.

Some of us practiced the “Bus Factor” as a joke. But, many a truth is spoken in jest. In 2006, my first company was acquired by a Startup that got transplanted from Silicon Valley to India.

We were not Linuses, but as an unspoken/unwritten pattern, we naively followed “if you get into an accident on the Mumbai–Pune Expressway.”[1] We never traveled together as we gradually relocated to Pune while still unable to give up Mumbai that quickly.

1. https://en.wikipedia.org/wiki/Mumbai–Pune_Expressway