Summary: L4-Hurd meeting in Ka

[Wolfgang Jaehrling]

Introduction
=-=-=-=-=-=-

Just a random weekend in August?  Well, not quite.  Several Hurd
people have been meeting in Karlsruhe.  Even better, we've been
meeting some L4 developers as well.  Here is a summary, in the form of
a travellers diary.  But first of all, a list of the people involved.

Hurd folks:
  Marcus Brinkmann (Germany)     Marco Gerards (Netherlands)
  Niels Möller (Sweden)          Moritz Schulte (Germany)
  Peter De Schrijver (Belgium)   Daniel Wagner (Switzerland)
  Wolfgang Jährling (Germany)

L4 folks:
  Espen Skoglund
  Uwe Dannowski
  Joshua LeVasseur


Friday, August 22
=-=-=-=-=-=-=-=-=

My train is 15 minutes late, I arrive at 11:20 pm.  As soon as I leave
the train, I already notice how the air smells for L4.  This is
Karlsruhe, a beautiful city with many fountains.  Daniel Wagner is
waiting for me at the train station as he arrived 10 minutes earlier.
We are walking to our hotel, which is not far away.  At the time we
enter one of our rooms in the fifth floor[*], the others are not here
yet.

[*] They've got an elevator, but I always walk the stairs because I
    don't like such modern technology.

Half an hour later, Marcus, Peter, Marco and Moritz (who is wearing
his blinkenlight all the time) arrive as well.  Soon we leave the
hotel to get some food, returning rather late at around 3:00 am.  The
main topics discussed at this evening were hardware drivers (though
not in the context of L4-Hurd), virtual memory management on L4-Hurd
and, of course, ``C vs. C++''.  (I think our most important conclusion
being that if a project takes more than 8 hours to compile when
written in C, and compiling C++ is really three times slower, you
cannot do daily snapshots anymore.)


Saturday, August 23
=-=-=-=-=-=-=-=-=-=

I wake up at 8:00 am, finish reading the small paper Marcus wrote
about porting the Hurd to L4 and enjoy listening to the trams passing
by in front of the hotel (as they sound almost like the music I
usually listen to).  At 9:15 am, Daniel and Marcus are waking up, but
not standing up until 9:30 am.  We take breakfast with Marco, Peter
and Moritz, reach the university at 11:20 am, where we meet Espen, Uwe
and Joshua.

Marcus is giving a talk about porting the Hurd to L4 - comparing the
design ideas of the Hurd and L4, explaining what is missing on L4 that
Mach provides to us and describing how we are going to approach these
issues and presenting the current state: The virtual memory management
has been worked on by Neal Walfield, Marcus designed a capability
system to do the IPC and actually started implementing it, Daniel and
Peter have started designing the device driver framework.  Marcus goes
on to talk in some detail about various problems he discovered and
possible design improvements in the Hurd that are desirable or
necessary on L4.  We also get many questions answered by the L4
people, especially Espen.

We fetch chinese food and Marcus is buying socks, since he forgot his
ones at home (next to the toothpaste).  We return to the university,
now Daniel and Peter show the current state of the device driver
stuff, receiving a lot of valuable feedback.  Marcus later gives a
wonderful talk for Marco, Moritz and me about the capability system
while Peter and Daniel are working together on the driver issue.

At about 9:00 pm, we meet Niels Möller and go to the ``Littfaß'', a
great restaurant recommended by Daniel (If you will ever be there,
don't forget to try the toilets!).  After that, we go into an Irish
Pub.  When we arrive at the hotel at 1:30 am, we are somewhat tired...


Sunday, August 24
=-=-=-=-=-=-=-=-=

Again I'm standing up at 8:00 am.  I start into the day by doing
something for the community (taking a shower).  At 9:50 am, I wake up
Marcus and Daniel (using water and brute force because everything else
fails).  We arrive at the university at 11:30 am, and start with a
talk about the task server given by Niels, including a lot of
discussion with Marcus, metamorphosing into a collection of small
groups talking about various stuff: Marcus presents his capability
system to Niels and Espen (today the only L4 hacker around), Daniel
and Peter continue with working on driver stuff, Moritz and me talking
about LISP implementation details and Emacs configuration, etc.
Several of us also discuss a way of making GNU/Hurd more secure by
giving users the option to only follow translators set by certain
users.  So the day passes with us being more or less productive.

Somewhen on the late afternoon, we leave the university and go to
another restaurant (actually more like a beer garden) where we
continue discussing a bit, then directly go back to Peters car at the
hotel (passing Niels hotel at 7:50 pm where Niels, Daniel and Espen
are leaving us).  Due to bad train connections on sundays and me
losing my ticket when running to the train so I have to get a new one
and take a later train, I arrive at home at 0:40 am.


Thanks to the L4 guys for providing us with a room and being very
supportive.  According to Marcus, it was definitely not the last Hurd
hacker meeting, and I am certainly looking forward to the next one.

Cheers,
GNU/Wolfgang

-- 
"A good programming language should have features that make the kind
of people who use the phrase `software engineering' shake their heads
disapprovingly." -- Paul Graham