[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Fwd: looking back at 2015

From: Robert Millan
Subject: Fwd: looking back at 2015
Date: Sat, 16 Jan 2016 12:45:48 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.5.1

FYI, some Hurd mention in Rumpkernel's yearly report

-------- Missatge reenviat --------
Assumpte: looking back at 2015
Data: Fri, 15 Jan 2016 15:30:50 +0000
De: Antti Kantee <pooka@rumpkernel.org>
Respon a: pooka@rumpkernel.org
A: rumpkernel-users <rumpkernel-users@freelists.org>

Hello, and a belated welcome to 2016,

As people who've been on the list for a while know, the visions &
roadmap mail comes during summer, simply because it's nicer to think
about those things while sitting outside enjoying sunny weather.  In
this mail we look back at 2015, if only to give an excuse to write light
prose for Friday afternoon delight.

For the rump kernel project, 2015 was undoubtedly dominated by the
unikernel use case.  Thinking back, at the beginning of 2015 only Martin
and myself were regularly contributing to the Rumprun unikernel (*).
Now we have at least a dozen people contributing more or less regularly
(**), especially as maintainers for rumprun-packages.  What has been
especially nice to see is that folks don't only contribute new packages,
they also stick around and maintain them.  Those who've been following
my rants know that I think making something maintainable and maintaining
it is much more demanding than just whipping something up and forgetting
about it.  It's also harder to find motivation for maintenance, at least
unless you have non-academic real world interest in keeping the subject
matter afloat.  A big thanks goes to the contributors of the packaging

*) back then, the Rumprun repository as we know it now didn't even
exist, it still existed as two separate repositories named rumprun-xen
and rumprun-baremetal.  In fact, for all I remember, we didn't even
market those as unikernels back then.  Their unification surely improved
my mental stability, and starting to call the result a unikernel also
shows how far unikernels came as a global concept in the past year.

**) As of writing this, the rumpkernel project on github has 23 members,
which means that 23 internet beings -- most of them probably human --
have push access to at least one repository under repo.rumpkernel.org

Now, while unikernels are the proverbial Soviet space pencil, if you
want to create a hardcopy of War and Peace, you'll most likely want to
reach for some other tool.  The folks who brought balance to the Source
by improving a different type of tool was the HURD microkernel project.
 They became, to the best of my knowledge, the first folks outside of
rumpkernel.org to incorporate PCI device drivers provided by rump
kernels into their system.  Again, it was not idle academic curiosity,
but rather solving a real problem for their system.  Nicely done!

On the documentation front, the wiki received improvements from a wide
number of people.  A few even contributed brand new articles.  Writing
documentation is harder than writing code, since it has to be written
against non-exact and not-well-defined human parsers.  That difficulty
is why it is very important to get a breadth of people contributing to
the docs.  Many thanks go to those who have chipped in on that front.

If you think there was some other significant type of 2015 happening
which failed to surf my brain waves, please do share.

This year, unikernels will probably go even further.  And I will be
sitting here as the grey lord (***) making sure things are in balance
for all uses of rump kernels.  Please keep pouring them in.

  - antti

***) knowledge of Dungeon Master is also a definite requirement for
literacy on this list.  wonder if that should be in the FAQ.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]