Current location

narf Source control manager Git

diff options
authorOlivier Mehani <>2018-01-24 11:27:17 +1100
committerOlivier Mehani <>2018-01-24 11:27:17 +1100
commit7de43f684d1ff03aa25aa385e10a66bb82d27e10 (patch)
parentcea1d2e8e281e17b95be59ccb3a119ed1e858b90 (diff)
Johnathan Corbet's talk
Signed-off-by: Olivier Mehani <>
-rw-r--r--2018lca/IMG_20180124_112447.jpgbin0 -> 74457 bytes
2 files changed, 108 insertions, 0 deletions
diff --git a/2018lca/2018lca.tex b/2018lca/2018lca.tex
index d92b63f..e4e9f96 100644
--- a/2018lca/2018lca.tex
+++ b/2018lca/2018lca.tex
@@ -698,6 +698,114 @@ Why would one need access to a bio hacking lab?
+\subsection{Keynote: Karen Sandler}
+ \item Pacemaker in 2012: Doesn't know what software it runs $\rightarrow$ best
+ printer story ever
+ \item Did not get the source code
+ \item Software Freedom Conservancy: Git, HomeBrew, \ldots
+ \item Still no source code
+ \begin{itemize}
+ \item Obtained DMCA exception for medical device research
+ \item Got called a conspiracy theorist by her electro-physiologist
+ \item Can communicate wirelessly (and deliver 830V)
+ \item Johnson \& Johnson had an issue with a remotely exploitable insulin
+ pump; they were transparent about it, and worked on fixing the issue
+ \item Shift from blanket trust in FDA approval (FDA doesn't review source
+ code)
+ \item FDA underresourced, can't review sourc code, but would be happy if
+ others could
+ \end{itemize}
+ \item Bug: while pregnant and anxious, heartrate increased $\rightarrow$ got shocked by the defibrilator
+ \item Companies become comfortable with the rules of the product if they need
+ it
+ \item Short-term corporate interests are not always aligned with long term
+ community ones
+ \item Everybody like to think of themselves as a good person
+ \item Don't give away control, it's harder to regain
+ \item Free Software is the future we are building for uor future cyborg selves
+\subsection{History rhyming in the kernel community and beyond --- Jonathan Corbet}
+ \item History repeats itself
+ \item Use-after-free bug in TCP fixed in 2007 found in similar code in DCCP
+ \item Bugs are like mushrooms: If you find one, you need to look around for
+ more --- Al Viro
+ \item Nobody is looking
+ \item Kernel changesets and releases have become very predictable
+ \item Android 8.0.0: 3.10 kernel (2013), 320,000 patches behind the mainline
+ \item Back in pre-2000: no source-code management, tracking patches was
+ painful
+ \begin{itemize}
+ \item none of the SCM tools were good enough at the time
+ \item until BitKeeper came round, designed for a distributed development
+ model
+ \item most of the problems disappeared in a few months
+ \item but not Free Software; number of annoying terms (Free for public
+ projects); ultimately the license was withdrawn in 2005
+ \end{itemize}
+ \item Lessons learned from BitKeeper
+ \begin{itemize}
+ \item Tools matter
+ \item Licenses matter
+ \item Tools can be made to work with the kernel dev model
+ \item Free-as-in-beer isn't sustainable
+ \item \ldots
+ \end{itemize}
+ \item Why are we still using email rather than GitHub? Is this like the
+ resistance to SCM in pre-2000? Is this conservatism impeding progress?
+ Should we be considering a proprietary service? GitHub's license is very similar to
+ BitKeeper.
+ \item Need to develop a new tool?
+ \item (Roughly half the audience runs their own email server)
+ \begin{itemize}
+ \item Getting email delivered from there to big operators is getting harder
+ \item Kernel developers needs to get GMail accounts so their PRs go
+ through
+ \item How Free is email?
+ \end{itemize}
+ \item Current hardware really is a stack of proprietary software on top of
+ which we run Free software (\eg, F00F, Minix 3 in ME, Meldown, Spectre)
+ \begin{itemize}
+ \item We need more work on Open Hardware
+ \item Building an Open CPU is costly
+ \item \ldots but so was building a Free OS in 1983
+ \end{itemize}
+ \item Copyright as a control mechanisms
+ \begin{itemize}
+ \item Google Play Services
+ \item Keeping control over what's being shipped with Android phones
+ \item Prevent vendors from shipping a phone with, \eg, Lineage OS
+ \end{itemize}
+ \item Platforms: everyone who tried to control one ended up failing
+ \begin{itemize}
+ \item IBM
+ \item DEC
+ \item SunOS/Ultrix/Irix/HPUX/AIX\ldots Windows
+ \item \ldots
+ \item Linux!
+ \end{itemize}
+ \item We set up to change the world, and we won.
+ \begin{itemize}
+ \item We have control over our computing systems
+ \item Not complete or perfect
+ \item But shows progress is achievable
+ \end{itemize}
+ \item But could we be blindsided (like DEC believing Unix wouldn't pick up, or
+ Sun / Linux); what could replace Linux?
+ \begin{itemize}
+ \item a smaller and nimbler kernel?
+ \item a more permissive license?
+ \item under some company's control?
+ \end{itemize}
+ \item History suggests that Linux is not forever
diff --git a/2018lca/IMG_20180124_112447.jpg b/2018lca/IMG_20180124_112447.jpg
new file mode 100644
index 0000000..ae116d8
--- /dev/null
+++ b/2018lca/IMG_20180124_112447.jpg
Binary files differ