Archive for December 2004
I just released 6.8.0-r4 into testing yesterday, two days before I go on vacation. Now that takes planning.
[old] Docs install to /usr/share/doc/xorg-x11-6.8.0
[old] Man pages install to /usr/share/man
[old] Fonts install to /usr/share/fonts
[new] /usr/X11R6/lib moves to /usr/lib
So, what’s left in /usr/X11R6, you ask? bin/, include/ and a few backward compatibility symlinks. bin/ will be the next to move — it probably will happen in 6.8.2.
I just got my Christmas presents from my mom. I sent her this huge list of books from Amazon, expecting her to get a couple. She bought all of them!
Fortuitously, I overslept today so I was here when the UPS guy arrived.
She got me “The Art of Assembly Language” and “Writing Great Code, Volume I: Understanding the Machine,” both by Randall Hyde. I’m excited to start diving into those.
Also, I’m fully stocked on quantum chemistry — seven new books ranging from symmetry and spectroscopy up through the basic dogma, electronic structure theory and computational chemistry.
Unfortunately, I won’t be able to make it to church Christmas Eve, as I do almost every year, because we’ll be driving up to Portland to stay the night before flying to Minnesota early Christmas Day. That week off will definitely be nice, though. I’m getting a little broken down. I’ve been missing a day of work/school about every two weeks, because I’ve just been overworking myself.
Hope everyone else enjoys the break as much as I will!
Oh by the way, I finally found a distribution of MOPAC7 that works. I needed to hack the Makefile a little once it was generated to work around problems with MAIN__ and main() in libf2c so I could actually get a binary. But other than that, it’s been great!
A couple of possible ways to “fix” the MAIN__ problem:
1. Add this to your C source:
int MAIN__( )
2. Add this to LDFLAGS:
-Xlinker -defsym -Xlinker MAIN__=main
Incidentally, I got Tinker working too. Apparently Firefox sometimes doesn’t like downloading complete files. wget –passive-ftp worked perfectly.
So now I’ve got a quasi-complete computational chemistry workstation here. Just gotta get some molecular dynamics stuff on here — already got quantum covered, as far as free programs go.
Reading Havoc’s blog reminded me of how I wanted to focus on usability, primarily to help in the creation of a Gentoo installer. It’s a little bit of a weird project, because it has two distinct target audiences: automated deployment and (clueless?) end users who want/need a GUI installer. So it’s been interesting.
The code is getting near finishing up a working alpha — but we’re looking for experts in OO and Python to help look it over and chip in some ideas, answer some questions and maybe write some code. Stop by #gentoo-installer on Freenode IRC if you’re interested. Experts on UI design are also welcome.
I read another quick thing on usability the other day, perhaps linked from gnomedesktop.org?
It also reminded me of my need to actually read Joel Smolsky's “User Interface Design for Programmers.” I’ve owned it for months, but it’s still a ways down on the “when I have free time” list.
I got a preliminary ebuild committed Saturday for 6.8.2RC1, plus our usual patches. Also, I added Benjamin Herrenschmidt’s Radeon PPC patches and Alex Deucher’s adaptation of the FBDev patch to Rage128. Recently I was made aware that PPC64 builds are broken with 2.6 kernel headers, but that patch breaks standard PPC and perhaps other architectures.
I’ve been diving back into computational chemistry programs in the last couple of days. There should be some sort of law that says all scientists must hire a real programmer to maintain their programs. I’m trying to get a couple of programs working, called MOPAC and Tinker — both of which are in Fortran, of course.
I know MOPAC used to compile a couple of years ago, but no longer — and nobody’s updated it, and nobody’s fixed it that I can find. The only distribution I’ve found besides the public domain stuff is FreeBSD, and MOPAC’s marked broken on >=5.0.
As for Tinker, well … I can’t even untar the source. My system is broken somehow, and I have no clue what the cause is. But the source is all up on its homepage, and the build process is incredibly obscure.
Not only did I pass my classes, I actually got respectable grades. I’m shocked. I thought I was going to fail out and go home. A in biochemistry, B in biophysics and A in research.
Not having a Pegasos is killing me. I’ve got that “last kid on the block” syndrome — it feels like everyone else already has theirs (probably because it’s true). Both Seemant and Stuart have mentioned how fast they compile, which would be really great since X takes forever.
There are some pretty cool possibilities that I’d like to try out, such as setting up a distcc cluster between the Pegasos and my x86 box with cross-compiling toolchains. Gentoo’s cross-compiling is really getting off the ground, so this would be a great time to help out with it if I could.
I’d also like to see whether there’s anything I can do to make them install and run smoother before X even gets involved. When I used the 2004.1 Gentoo CD, it was a little bit hackish. I had to type in a long command line at the firmware prompt to boot the CD. I’ve heard 2004.3 was an improvement, but I didn’t get a chance to try it out. Also, I put Gentoo into the boot menu rather than having to type things in at the firmware prompt, which seemed like the current status quo. Getting from the point of pressing the power button to an X login screen with little to no effort is a crucial part of being a usable Linux desktop.
With rights come responsibilities. Remember to think about the consequences of your actions, even if you’re fully within your rights to do them.
I’m really looking forward to getting a working Pegasos. I’ll have time for it between Wednesday and Christmas, so it would be great to get it soon.
I’m about to email Genesi and ask for a status update.