Conversation

Jonathan Corbet

20 Years ago: the BitKeeper license changed, making it unavailable for kernel development.

https://lwn.net/Articles/130746/

It drove home the perils of relying on proprietary software and spurred the creation of Git - a significant event, overall.
4
65
76

@corbet Also a great read, how BitKeeper was "reverse engineered":

https://lwn.net/Articles/132938/

0
0
0

@corbet I was around then, there were some exciting times of experimentation in version control before git came to dominate, influenced heavily it seems by the Linux kernel choosing git, and perhaps not unrelated that Torvalds wrote both.

Darcs was considered for a minute for the Linux kernel, but was too slow on large tree operations.

1
0
0

@corbet yeah, the master's tools cannot dismantle the master's house and all that. If not for git, foss collaboration would have been so, so much worse today

0
0
0

@markstos
AFAIR, Torvalds wrote git precisely *because* the existing alternatives wouldn't cut it for the Linux kernel...
@corbet

1
0
0

@mnalis @corbet Agreed, but there was a sense at the time that `git` might be a temporary choice until something better came along.

Git worked very well for the scale of the Linux kernel.

0
0
0

@corbet and Andrew Tridgell presented an entertaining talk about getting metadata from bitkeeper at linux.conf.au

0
0
0