Writing about L4 and the future Information Society

[Gernot Heiser]

Ho John,

Interesting article. Looks like you've had some truly traumatic
experiences.

I'm not sure what to comment on, as you aren't really asking any
questions or make controversial statements. All I can comment on is:

>>>>> On Sat, 28 Jul 2007 11:54:57 -0400, "John van V." <john.van.v at gmail.com> said:
JvV> [...]

JvV> In 2005, when I was about to start
JvV> investing in my own truck fleet, I had an epiphany -- to teach middle school
JvV> kids science, and especially technology.  I believe that middle school and high
JvV> school students, from my experiences, are the best potential technology core
JvV> from the perspective of the future-- and they work for free.

The importance of investing into school kids cannot be underestimated,
although I see them as the future technologist, not a free work force.

Particularly in the present situation where in many industrialised
countries kids are losing interest in technology and science. If this
trend continues we will create a society with a very strong class
structure, where the majority has no understanding of technology and
limited access to its benefits, and even less understanding of its
dangers. Needless to say, this is also a big threat for the economic
competitiveness of those countries.

I have been tangentially involved with outreach activities at
UNSW. And the scary thing is that in high school, it is essentially
too late. The kids have already decided whether they are interested in
science and maths, and those who have decided they don't like it are
essentially already a lost cause. Regrettably, this is particularly
true with girls, who at this age are very sensitive to peer pressure,
and are being told by their peers that maths isn't a girl thing. In
Australia, there are also studies showing that high school career
advisers are also discouraging girls from science and
engineering.

Essentially, the battle is already lost in high school, it has to
start earlier. Our experiences with running workshop for year-five
pupils (especially girls) are much better, they are still open at this
stage. This is the time where the interest needs to be nurtured.

JvV> Also, I should mention that many people are saying that OS development in
JvV> general has slowed to a near halt.  This appears to me to be so, though all of
JvV> you involved in L4 are not to blame!  I feel that L4, as the leading new OS, can
JvV> be a focus for, or perhaps locus of, all the things we need to do, and all the
JvV> reflections we need to make based on the nearly two decades of successful, yet
JvV> ultimately heartbreaking experiences, in the new Information Society.

The observation that "systems research is irrelevant" has been made
before by Rob Pike (http://herpolhode.com/rob/utah2000.pdf) and
indeed, OS research was by many considered dead a few years ago. This
isn't the situation at the moment, though. For one, Linux has changed
the game by making OS code much more accessible, students can now
again do research on real systems. Then virtualisation has created a
lot of interest and activity in OS issues. However, Rob's observation
that people are still largely using the same 40-year-old technology is
still true (and virtualisation is essentially used to hack around the
limitations of broken operating systems).

The reality, though, is that some of this is the inevitable result of
the commoditisation of PCs, and the resulting huge inertia in the
basic architecture, processor as well as OS. I don't think there's
much hope in changing the PC world in the foreseeable future.

Embedded systems, however, are a different ball game. The embedded
systems industry, in different verticals at different times, is
realising that they have reached the use-by date of their RTOS
technology. Hence they are forced into a change of OS technology, and
this is the chance to put in something that's good. I sure believe
that this is L4, and that's the reason we have set up Open Kernel
Labs.

If you are interested in where we believe things are heading, there is
a paper in this month's Operating Systems Review.

Gernot