Core work still going on 33 years later [LWN.net]

Core work still going on 33 years later

Posted Oct 23, 2024 4:38 UTC (Wed) by raven667 (guest, #5198)
In reply to: Core work still going on 33 years later by willy
Parent article: The long road to lazy preemption

I think this is right, Linux is maybe unique in how aggressively it's been refactored and re-imagined internally over the years, most projects avoid rework and try to achieve stability by not changing things, but not Linux.

What I was thinking is (aside from the tty layer that no one wants to touch with 2m borrowed pole) how many distinct Linux kernel designs have existed over the last 30+ years? What would define the eras, since change is happening all over, removal of BKL, switch from stable/dev branch in 2.6 to continuous integration, udev, some particular scheduler or memory allocator? What would a kernel developer see as distinct coherent design eras? How much code has been unchanged in the last 5y, 10y and again the 5y, 10y before that, how many Ships of Theseus have been built?

to post comments

Core work still going on 33 years later

Posted Oct 23, 2024 13:36 UTC (Wed) by raven667 (guest, #5198) [Link] (2 responses)

Just because I liked the analogy, the Linux kernel has started from a wooden longship and been refactored into a full sailing galeon and then further refactored into a modern container ship, changing scope and structure along the way.

Core work still going on 33 years later

Posted Oct 23, 2024 15:16 UTC (Wed) by Wol (subscriber, #4433) [Link]

I think it started as a coracle, or dugout canoe ... probably coracle, that feels more finnish :-)

Cheers,
Wol

Core work still going on 33 years later

Posted Oct 31, 2024 10:34 UTC (Thu) by FluffyFox (guest, #162692) [Link]

This reminds me little history of the "very first version" of Linux which initially started as a program doing multitasking with 2 threads and today it become a versatile and flexible kernel