Interview by David Ford with Jonathan Walther, on fantasy, literature, life and
primarily Xouvert the branch of Xfree86, which he is leading.
PART ONE (Sunday August 17, 2003)
DF: Hello and thanks for giving me this interview, would you mind saying a
little bout yourself?
JW: Hi. It’s nice to be here. I’ve been a Debian developer for more than 5
years, and am a proud father of two. I enjoy making organic all-grain beer at
home. My friends tell me it is delicious. I agree with them, of course.
DF: Sounds wonderful, might I have a glass?
JW: If you’re in the neighborhood, definitely. My latest batch is called “Owen
Meany Bitter” and should be ready to drink in a week.
DF: So what has been your involvement with XFree86 prior to this branching?
JW: Interested observer, daily user, and increasingly, a frustrated user.
Xmame started reliably crashing X on me; the OpenGL acceleration stopped
working for my Matrox G400 video card. Definitely a bit of frustration.
JW: As an X developer, I contributed to the ratpoison window manager with some
minor patches and the ratmenu value-added utility. I worked on some Xlib apps
with Bowie Poag to test some of his incredible new user interface ideas.
JW: The scrollball was a particularly fascinating UI idea from Bowie that we
coded and tested two years ago. Recently he has come up with “pogo”, a sort of
super-glitzy looking dock.
DF: Can you tell me more about scrollball and pogo?
JW: Scrollball and pogo are totally unrelated to Xouvert and XFree86. Perhaps
we could discuss them another time?
DF: That might be a good idea, they sound interesting so aside from
frustrations, what lead you to consider branching such a big project?
JW: A few months ago I noticed that XFree86 had made more public its emphasis
on stability, as opposed to new features. So I saw a need for a development
tree that wasn’t as worried about trying wild new ideas. I also saw that arch
had made such a project possible, without compromising stability too much. It
seemed the only way to convince people that it was possible was to actually go
out and do it.
DF: Tell me a little about Arch then, it seems to have an important
relationship with the new Xouvert project.
JW: Arch is a project I’ve been watching for the last year. Its author, Tom
Lord, has made all the right development moves. He prototyped it in shell
script to prove the concept, then rewrote it in C for speed. He wrote it to
handle things that CVS and Subversion just cannot handle, by their designs. But
most especially I liked the way arch allows you to easily make local
modifications to a source tree, then keep updating your source tree from the
original source, but WITHOUT LOSING YOUR MODIFICATIONS. This was what Xouvert
needed to be able to be a branch of XFree86, and NOT a fork.
DF: Exciting stuff. How much ‘setting up stuff’ is left to be done?
JW: As far as getting to a point where people can start coding, I am finishing
a tagging script that prepares the CVS sources to be compared to the current
arch sources. After that, it will be a simple import and make-mirror to the
Savannah web server. The IRC logger is working, but I need to set up a cron
script that will process the raw logs into easily accessible HTML, on a yearly,
monthly, and daily basis. Finally, we want to get a wiki going again, but with
a little more security so we can deal with vandals as they arise. Today’s
episode with our wiki was rather embarrassing.
DF: I think I missed that, anything serious?
JW: A vandal put an ASCII graphic of goatse.cx on our wiki, and had a bot
refreshing it every 2 minutes. So we couldn’t change it back, noah and mharris
of the Xwin project dealt with the problem as soon as they became aware of it,
so it isn’t a problem right now, but we are going to wait a bit and see what
comes down the pike in the way of Wiki space for the Xouvert project.
DF: Sorry to hear you had troubles, would you please tell me a little bout the
plan and purpose behind project Xouvert?
JW: Xouvert is intended to be a project for finding new talent to contribute to
XFree86, and for pushing the envelope of what is possible with todays hardware.
Bringing tomorrows technology to your desktop today is our motto for a reason.
DF: Any idea of the sort of changes that will be made?
JW: There may be some directory restructuring in the source to enhance
modularity. We definitely want to make individual video drivers independently
available on a daily basis, so people can always have the latest and greatest
drivers for their video cards. We plan to try out some things like enabling
backing store by default, supporting 3d stereo vision goggles, and the like.
We definitely want to see what we can do about integrating the DRI/DRM project
into the tree. As well as other relevant projects, like NX (compression that
works over 9600 baud modems) and of course we are interested in the Xr/Xc
projects, and the Xft project. We also plan to more fully support resizing the
virtual size of the screen on the fly, something that windows users have had
DF: So who is Xouvert’s target audience?
JW: Right now, we are interested in developers with coding abilities and ideas
for the future of X. But we have a second target audience: all users of X11.
Knoppix has some very nice, effective setup scripts, which we may be able to
DF: Well its 5 am here and I’m tired, mind if we call it a day and return to
JW: Not at all, get some sleep. I look forward to completing this interview
with you. have a good night!
PART TWO (Monday, August 18, 2003)
DF: Good morning. Sorry for being late in returning to our discussion today I
was working on my fantasy novel.
JW: I used to read a lot of fantasy and science fiction. Haven’t had much time
DW: I’m working on a little Fantasy/Psychological Horror novel, the plots
worked out long ago, just attempting the first write up.
JW: Nice. When I was younger, the Lord of the Rings trilogy really swept me
off my feet. More recently, the Illuminatus! trilogy made me feel like I knew
what it was like to have an acid trip.
DF: I haven’t read the Illuminatus! trilogy, if its that good I’ll have to.
JW: Yes, the Illuminatus! trilogy really blew my mind. The facts were all
false, but despite that, the book was truer to reality than any other book I’ve
DF: What’s the plot?
JW: It has so MANY plots… I don’t know where to start in describing it.
DF: I’ll have to get a copy when I finish this book and (hopefully) make a
small fortune. Until then I’m poor.
JW: Yeah, or get it from the library, it’s a classic, written about 20 years
after Lord of the Rings.
DF: That is an idea, been forever since I visited the library. My local one
doesn’t have anything good there and it takes them a fortnight to order it in.
And I lost interest because the bookshop around the corner has better books in
stock, and can order them in less time. So I’m sitting here with the guy who
would implement X16 a decade before we’re done with X11 and I’m discussing
JW: Everything is a fantasy of one sort or another. But some fantasies actually
have a chance at becoming real. Heck, after a couple years, Xouvert may
participate in standards efforts to design “X12”.
DF: There is the question of whether it matters.
JW: To people who want smooth, fast desktops, it matters. To people who want
glitzy graphics and nice looking/acting/feeling games, it matters.
DF: I mean with open X11 and BSD style licensees your pretty much setting an
open standard with or without a standards body.
JW: So when the day comes, Xouvert will be ‘the X12’ implementation. But first
we need to figure out what X12 is. It won’t be a radical change to X11, just a
rebalancing. In fact, the X11 protocol, with its extensions, is so flexible
that I suspect we won’t need to jettison backward support for X11 at all; all
that’s needed is to add “extensions” one piece at a time, and rewrite our core
apps to use them, and not use the deprecated features. Then over time, other
apps can get rewritten to use the newer, whizzier extensions and abandon the
crufty old ways of coding to X11 protocol. X11 is a general enough protocol,
like TCP/IP, that you can invent a web server or ftp server to run over it,
without having to change the protocol. Hence, backwards compatibility never
need be an issue as we move forward into the future.
DF: something needs to be done with it though if open source wants to compete
on the desktop, OS X is beautiful to behold already, and even Microsoft seems
to be ahead at the moment from the reports of longhorns new graphics core the
video’s suggest its very fast and powerful.
JW: I think the KDE and GNOME teams are working very hard to make a desktop
that not only competes, but far outstrips OS X and Aqua. Smaller teams and
individuals are also privately working to get us toward that goal. “Pogo” is a
good example, you can check it out on Savannah and Freshmeat. As KDE and GNOME
find things they need in the X server, Xouvert will be here to stand and assist
them. Neither of us working alone can achieve a beautiful, modern desktop. It
takes all of us.
DF: You’re fun to talk with, what do you like to be called?
DF: A pleasure meeting you Jonathan.
JW: Likewise it has been enjoyable meeting you. Our website has gotten more
than 90,000 unique hits from slashdot over the past few days.
DF: And OSNews and other places. Who is hosting you? It must be expensive.
JW: I’m not sure what the hosting cost is. Webspace is donated by a very good
friend of mine, John Sokol. John Sokol was one of the original 386BSD
developers, before it was branched to become FreeBSD and NetBSD. It is a sad
thing that all his credits in the code got taken out, over the years.
DF: Its sad that *BSD is as low profile as it is.
JW: John continues to do amazing things with BSD, especially in the area of
video streaming and compression. Being Canadian, I supported OpenBSD just to be
patriotic. But after I used it a bit, I realized it is actually much nicer than
the other distributions of BSD and Linux. It is small, tight, and has no bloat.
The security is what really sold me on OpenBSD though.
Interview conducted by David Ford, a freelance writer. He is looking for work
while also creating a chat empire over at P10Link.net.