linux.conf.au
Leaving now for linux.conf.au,
hopefully I’ll reappear on the other side in 30 hours time.
(This post was originally found at http://log.ometer.com/2004–01.html#10)
Leaving now for linux.conf.au,
hopefully I’ll reappear on the other side in 30 hours time.
(This post was originally found at http://log.ometer.com/2004–01.html#10)
Frederic,
indeed, thanks for the desktop-file-utils hacking. And yeah, sorry
about the code quality. A quick hack mutated out of
control…
Interesting discussion
of activation continues on message-bus-list, see also the earlier
posts and an
older thread.
Seth Nickell will be in Raleigh for orientation tomorrow. Due to
amazingly fast relocation skills, he’s the first person to start work
as part of our recent
hiring. Jobs remain available — there’s another spot for an
interaction designer, if you are one.
(This post was originally found at http://log.ometer.com/2004–01.html#5)
Let’s see if my custom weblog hack can handle multiple years…
Need help and volunteers for the D‑BUS todo
list, we need to get a lot of those things fixed before we can
ship D‑BUS in a release. There’s substantial work left, even
though people seem to be using D‑BUS left and right. Some high-level
tasks such as completing the GLib bindings aren’t even in the TODO.
Everyone’s sick of me mentioning it, but try “make check-coverage”
and then find exact untested lines of code with “decode-gcov
filename.c” — extending the test suite is easy and valuable, if
you want a simple place to start hacking D‑BUS. Laugh if you want but “make
check-coverage” rocks and we should have it for more modules, I deeply
regret that I didn’t do this for Metacity.
Ray Strode is kicking
ass fixing session management. Module “msm” in CVS has the work.
I’m hard at work on my linux.conf.au presentation, at least
the keynote one. I need a good title so people will get out of bed by
10am and come listen. Jeff has me signed up for something at the GNOME
miniconference as well but I haven’t figured out what that something
is. Plenty of time to think about it during the 30-hour marathon
flight to .au I suppose.
I believe I’ve officially failed (again) to get the new menu system in GNOME
2.6, because the deadline has more or less arrived. Even though the
code is in principle 90% complete and thanks to Heinrich Wendel has a
test suite. Sadly I’ve already moved cvs HEAD of the redhat-menus
package in Fedora to use the new spec, so for Fedora Core 2 we’ll have
to revert that or else backport the new menu system to GNOME 2.4.
If you’ve noticed a theme, I have less and less time for hacking; I
think what I’m doing instead at the moment will have more impact, but
time will tell.
We got a miniature dachsund a few days ago; he refuses to do anything
other than sit in one particular spot in the apartment. If you move him,
he runs back to the spot immediately via the shortest route. You have
to bring food to this spot or he just goes hungry rather than move.
(This post was originally found at http://log.ometer.com/2004–01.html#4)
Having trouble updating this page lately; way too much going
on, and the holidays combined with my trip to linux.conf.au are enforcing enough
downtime to make the uptime even more frantic.
Tonight’s Daily Show rerun had clips from MSNBC where they had
constructed a model of Saddam Hussein’s hole from plywood. The MSNBC
anchor climbed inside it to demonstrate… how to lie in a hole? Some
things can’t be explained, even by Jon Stewart.
I loved Lost in Translation.
Are we at some
critical point where Microsoft starts to lose big? I don’t agree
with many of the specific comments/rationales in that article, it is
the Inquirer after all. But coming into 2004, I do share a sense that
a corner is being turned, in part because Linux is moving beyond the
server, where UNIX was historically strong. The Linux desktop is
finally gathering real momentum.
If you count questionable
flamewars I’ve been following the Linux desktop daily since at
least 1997. At some point in there I started writing
some awful code and later on learned what I was doing, thanks to
copious cluebat from people such as Owen and Darin and Jon Trowbridge
and the many others who taught me to write software.
When I got involved I had no idea it would be six years of my
life before things really started to heat up. Matthias Ettrich started
KDE a year before I was even involved, and I faintly remember the
early GIMP
and GTK+. Can you believe it’s been so long? Many people have been
working far longer, of course.
I feel exhausted sometimes; but everything up to now has been
small! We’ve achieved perhaps 1% desktop marketshare, and the
resources and effort around the Linux server have dwarfed what we have
for the desktop. In the big picture we’ve barely gotten started.
Joel says good
software takes ten years.
The hardest and longest work is only beginning. No matter how delayed
it turns out to be, Longhorn will be here before we know it. Most of
the code we need to write hasn’t been written, and even as we face the
technical challenges we’ll have the organizational task of scaling the
community to accomodate many more users and developers.
Will Longhorn mean we move from HTML to XAML,
or will as many people move to Linux as move to Longhorn? How will C#
play out vs. Java? How will multimedia and DRM fit in? Will Microsoft
have to change course on open source as it did with respect to the
Internet? Will the open source movement be confined to the Linux
operating system, or as Red Hat is
hoping be a broader phenomenon? What will happen to Apple? What
legal and political challenges will we face? When will we meet
regular, nontechnical people on the street who take Linux for granted?
Will we ever see the long-predicted move away from the general purpose
desktop OS in favor of appliances and set top boxes? When will we go
from 1% marketshare to 2%, and to 5%?
It will be many years before we know if the original dream of a
widespread free software desktop will come true — that’s the scary
part. But the exciting part is that it seems on track so far, in fact
it seems to be catching fire. And there are countless opportunities for
anyone to dive in and have a huge impact on what our desktop is like,
and whether it succeeds.
(This post was originally found at http://log.ometer.com/2003–12.html#30.2)
PowerPoint and OpenOffice.org Impress really are a bit clunky. I want
to try Keynote. Was
Keynote really implemented by one guy in one
year? If so, why hasn’t someone cooked up a nice open source
alternative? Seems like a fun project to me.
(This post was originally found at http://log.ometer.com/2003–12.html#30)
BTW, if you apply for
a Red Hat job, you may want to send an extra copy of your resume
to hp@redhat.com.
I’d encourage applying as soon as you can if you’re interested, even
if you aren’t available to start immediately. We tend to fill
positions quickly.
(This post was originally found at http://log.ometer.com/2003–12.html#13)
My first memory of Ettore is at GUADEC in Paris, where people went
around yelling “Ettore!” because it was fun to say.
It’s not so fun to find out he’s gone. Deepest sympathies to all his
friends and family, and especially to the Ximian team.
(This post was originally found at http://log.ometer.com/2003–12.html#11)
If you’re graduating this year or just interested in something new,
you might consider these
available positions at Red Hat.
(This post was originally found at http://log.ometer.com/2003–12.html#10)
Some desktop developers at Red Hat have migrated to our Boston office
over the last couple of years. Plus some new hires have started out
there. Net result, it turns out everyone is in Boston and I’m in
Raleigh. So I decided to follow the trend and go where the action is.
This week we made a second trip to look for a place to live, and
finally succeeded despite a massive snowstorm. One less worry, though
something could still go wrong in theory.
The house we may soon own was built in 1890 and has a white picket
fence and an apple tree. We looked at a lot of houses and condos, and
the choices are old and classic or new and sort of blah. A couple of
new-ish houses that were nice, but not a lot. Mostly designed with a
view to bullet points instead of good taste.
I love modern houses, this
one is really nice for example, but these houses exist primarily
in magazines. The opposite extreme can be fun too though.
(This post was originally found at http://log.ometer.com/2003–12.html#6)
At the New York summit Nat brought up
collaboration repeatedly, and collaboration is the theme of the bounty contest.
A couple of flash demos: Kubi (skip
the boring intro) and
Groove
(hope you can survive the horrible music). Slashdot also just posted
this interview
about IM which is related.
Kubi and Groove have the idea of project teams that share resources
and collaboration forums, something I didn’t notice in the bounty
projects. What if we had a platform library that exposed your project
teams. Teams would be made up of people in your contacts
database. Then your IM client, Evolution, office apps, shared
whiteboard app, panel applets, etc. could all take advantage of this
team information. Perhaps there’s some concept of “project you’re
currently working on” and information relevant to that project would
appear — recently-used files, bookmarks, IM presence of team members,
related email. Perhaps the active project is autoswitched if you open
a document or mail related to a particular project, even. Somewhat dashboard-like in effect.
Here’s another angle to think about it: what if each mailing list
you’re on had associated chat channel, whiteboard(s), files, calendar,
and so forth. So the GNOME release team would be a project, for
example. To have a meeting, the team members just make release team
their active project, and they suddenly see a list of other present
project members in their panel, have a chat window open, have the
release schedule calendar open, have a shared whiteboard.
Better — what if the active project is per-workspace, so you can have
more than one at a time. Or what if we autocreated a workspace for
each active project, perhaps even replacing the idea of a workspace
with the idea of a project? The advantage of a project over a
workspace is that a project has a name and saved state (apps that are
open, documents that are related, etc.) and can be activated or
hibernated. We could also allow scratch projects with only one member
(for that extra workspace full of random clutter).
Maybe that’s not the right approach, but the point is we could
probably do cool stuff by giving the desktop the concept of a project
team.
A hard problem is how to make it easy for anyone (even Windows users)
to join a project. One approach is to have a special Windows
client. Another might be to use standards such as webdav or ical to
put information at URLs, and then email/IM the links to Windows users.
Another might be to have a server that exposed a web-based interface
to the project data.
A much simpler task than this grandiose projects plan — we were
wondering at the NY summit why nobody has implemented a shared
whiteboard app. Seems like it’d be pretty easy, perhaps using Jabber/XMPP.
(This post was originally found at http://log.ometer.com/2003–11.html#27)