| |||||||||||||
The Linux Foundation's kernel development paperThe Linux Foundation's kernel development paperPosted Apr 1, 2008 18:38 UTC (Tue) by ds2horner (subscriber, #13438)Parent article: The Linux Foundation's kernel development paper
I have a number of editorial comments (rewording for clarity, dual meaning of phrases, etc)-
should I post them here?
The most confusing to me (such that I originally thought it was a major internal constancy
within the document) is the units on figure 5 -
which on the surface appear to be "Lines of code" per "2.6.x kernel release" by "Added",
"Modified" and "Deleted".
The better y-axis is "Lines of code per day" for the "Rate of change" measure.
So where should I post other minor changes like -
"Averaging" vs "Summing up" in the following statement,
and philosophical/semantic/interpretation issues like -
disputing that the elapsed days between releases ("Integration Days"), should be termed "Days
of Development". (Most of the actual "development elapse time and effort" for any code
integrated in the next release is not within that window, but typically performed months
before.)
(Log in to post comments)
The Linux Foundation's kernel development paper Posted Apr 1, 2008 22:38 UTC (Tue) by ds2horner (subscriber, #13438) [Link] Why does Table 1 contain 2.9.11 and 12 when these predate the conference decision held July 2005? "At the 2005 Kernel Developer Summit in Ottawa, Canada, it was decided that kernel releases would happen every 2-3 months, with each release being a "major" release".
The Linux Foundation's kernel development paper Posted Apr 1, 2008 23:06 UTC (Tue) by corbet (editor, #1) [Link] That decision simply ratified the de-facto process which was already in place; what really happened in 2005 was the adoption of the merge window discipline. The paper goes back to 2.6.11 because that's when the git history starts. That's when we start to have really good information about where the code is coming from.
|
|||||||||||||