Installing OpenWRT is not for beginners. What follows is not a step-by-step guide to every keystroke, but rather an overview that gets you past the particular hurdles of getting OpenWRT installed on the router, getting it to boot off the hard drive, and installing BackupPC and its dependencies. Please know what you’re doing before you get started.

I bought:

1. Asus WL-500gP wireless router. It has two USB 2.0 ports in addition to easy JTAG access if I ever get in trouble.
2. An external Hitachi 640G USB2.0 hard drive.

Together, the two parts set me back about $160. I decided to install OpenWRT, a distribution of Linux designed for wireless routers, and BackupPC, software that automates the backups.

1. Install OpenWRT as normal. I recommend installing the latest generic openwrt-brcm-2.4-squashfs.trx firmware. For me, that was Kamikaze 8.09.1. It was necessary to use tftp to install the firmware.
2. Configure the router to pivot_root to the USB hard drive on boot. Follow the “How do I boot from the USB device (prep)” and “Boot Configuration (Kamikaze)” guides on this wiki page. You don’t need to do anything related to hotplug. The page is outdated when it comes to adding the script to /etc/init.d/rcS, just modify your existing file to include the first three lines of the code on the wiki.
   • My /etc/init.d/pivotroot:
     1. #!/bin/sh
     2. # change this to your boot partition
     3. boot_dev="/dev/scsi/host0/bus0/target0/lun0/part1"
     4. # install needed modules for usb and the ext3 filesystem
     5. # **NOTE** for usb2.0 replace "uhci" with "ehci-hcd"
     6. # **NOTE** for ohci chipsets replace "uhci" with "usb-ohci"
     7. for module in usbcore ehci-hcd scsi_mod sd_mod usb-storage jbd ext2 ext3 ; do {
     8.         insmod $module
     9. }; done
     10. # this may need to be higher if your disk is slow to initialize
     11. sleep 4s
     12. # mount the usb stick
     13. mount "$boot_dev" /mnt
     14. # if everything looks ok, do the pivot root
     15. [ -x /mnt/sbin/init ] && {
     16.         mount -o move /proc /mnt/proc && \
     17.         pivot_root /mnt /mnt/mnt && {
     18.                 mount -o move /mnt/dev /dev
     19.                 mount -o move /mnt/tmp /tmp
     20.                 mount -o move /mnt/jffs /jffs 2>&-
     21.                 mount -o move /mnt/sys /sys 2>&-
     22.         }
     23. }
   • My /etc/init.d/rcS:
     1. #!/bin/sh
     2. # Copyright (C) 2006 OpenWrt.org
     3. if test $2 == "boot" ; then
     4.   /etc/init.d/pivotroot
     5. fi
     6.  
     7. run_scripts() {
     8.  for i in /etc/rc.d/$1*; do
     9.   [ -x $i ] && $i $2 2>&1
     10.  done | $LOGGER
     11. }
     12.  
     13. LOGGER="cat"
     14. [ -x /usr/bin/logger ] && LOGGER="logger -s -p 6 -t sysinit"
     15.  
     16. if [ "$1" = "S" ]; then
     17.  run_scripts "$1" "$2" &
     18. else
     19.  run_scripts "$1" "$2"
     20. fi
3. Set up a swap file. You’re going to need the extra memory, especially when doing large batches with rsync. The basic shape of this is below for a 256M swap file. You might want to use the LuCI interface to make this permanent (or edit /etc/config/fstab).
   1. opkg install swap-utils
   2. dd if=/dev/zero of=/swap.file bs=1024 count=262144
   3. mkswap /swap.file
   4. swapon /swap.file
4. Tweak opkg to not install to flash anymore. Comment out option overlay_root /jffs from /etc/opkg.conf. Otherwise you’ll have several hundred gigabytes of free space, but opkg will still refuse to use it.
5. Install BackupPC dependencies from packages. You’re going to need a whole heap of samba and Perl.
   1. opkg update
   2. opkg install perl samba-client samba-server
   3. # Helpful symlink for perl
   4. ln -s /usr/bin/perl /bin/perl
   5. # Install all perlbase packages
   6. opkg list | grep -o -E perlbase-\\w+ | xargs opkg install
6. Download and unzip BackupPC
7. Install (yet more) BackupPC dependencies manually. Both CPAN and CPANPLUS are irretrievably broken on this platform, so run BackupPC-3.1.0/configure.pl and install what isn’t there by extracting the contents of /lib of whatever CPAN module you’re installing to /usr/lib/perl5/5.10.
   • If it asks you repeatedly to install Pod::Usage and you have already, type this in the shell: perl -MPod::Usage and figure out what’s still missing when it dies.
8. Install BackupPC. Finally configure.pl will start asking you questions and let you install. Go for it. You probably want to have it run as the root user.
9. Configure BackupPC to start on boot. The initializing scripts provided in BackupPC-3.1.0/init.d aren’t bad, and the one for Slackware is the best. Here’s what landed in my /etc/init.d/backuppc (Modified from the Slackware script). After copying this to your router, run /etc/init.d/backuppc enable to get the service to start automatically.
   1. #!/bin/sh /etc/rc.common
   2. # Startup init script for BackupPC for OpenWrt
   3.  
   4. START=95
   5. STOP=10
   6.  
   7. start() {
   8.     /usr/local/BackupPC/bin/BackupPC -d
   9. }
   10.  
   11. stop() {
   12.     /usr/bin/pkill -f "/usr/local/BackupPC/bin/BackupPC -d"
   13. }
   14.  
   15. restart() {
   16.     stop
   17.     start
   18. }
   19.  
   20. reload() {
   21.     /usr/bin/pkill -f -HUP "/usr/local/BackupPC/bin/BackupPC -d"
   22. }
10. Set up BackupPC admin interface. Here are some general steps to shoehorn it onto the router:
    1. Copy BackupPC-3.1.0/cgi-bin/BackupPC_Admin to /www/cgi-bin
    2. Edit /etc/BackupPC/config.pl to set $Conf{CgiAdminUsers}='*'; This option does eliminate password requirements on BackupPC, which is appropriate for my purposes. If it’s not for yours, configure Busybox’s httpd to set $REMOTE_USER.
    3. Copy .css and .js from BackupPC-3.1.0/conf and .png and .gif files from BackupPC-3.1.0/images to /www.
    4. Drop a link to the BackupPC admin interface in /www/index.html, and get rid of the meta refresh in the head of that file.
11. Configure BackupPC. This is up to you. BackupPC is a complex package that needs to be customized to your particular installation. Go wild!

Related posts:

1. Good directions on getting BackupPC working I recommen

From this, you get interesting, relatively easy to solve equations. For example, if you are in state P_0 (in the beginning of the game) then P_1 = T x P_0 (the matrix product). In fact, P_n = T x P_(n-1). That means that you can chain them together, so that P_n = T x T x … for n “Ts” … x T x P_0. ²

Most of the transition probabilities are just 1/6 (a fair spin of a six-valued spinner). ³ But when you hit a chute or ladder, your destination changes, which makes it much more interesting. ⁴ It also means that there are combinations of spins that could result in a game of infinite length.⁵

Some results:

1. According to the analytical approach, you should win in 7 turns 0.04% of the time. This is the fastest you can win, and disagrees with toolic’s empirical code significantly. Toolic estimated the probability at 0.15% (1,597 out of 1,000,000 games played). I don’t know where the discrepancy comes from.
2. There is a 50% chance that the game will end for a given player in 30 turns.
3. There is a 90% chance that the game will end for a given player in 68 turns.
4. The game does asymptotically converge to the end square, so while you could loop around forever, the chances of getting to turn 1000 are far lower than winning PowerBall several times.
5. If you rolled only 2, you would win in 30 moves. The chances of this happening are 1 in 10,737,41,824.
6. If you rolled only 4, you would win in 17 moves. The chances of this happening are 1 in 131,072.
7. If you rolled only 6, you would win in 20 moves. The chances of this happening are 1 in 1,048,576.
8. Rolling only 1, 3, or 5 results in infinite looping.

Here’s the perl code to do it. You’ll need to install Math::Matrix.

use strict;
use warnings;
use Math::Matrix;

my $matrix_size = 100;
my %transition = (
   # Ladders
   #  Start
         1  => {end =>  38, action => 'up ladder'},
         4  => {end =>  14, action => 'up ladder'},
         9  => {end =>  31, action => 'up ladder'},
        21  => {end =>  42, action => 'up ladder'},
        28  => {end =>  84, action => 'up ladder'},
        36  => {end =>  44, action => 'up ladder'},
        51  => {end =>  67, action => 'up ladder'},
        71  => {end =>  91, action => 'up ladder'},
        80  => {end => 100, action => 'up ladder'},
   # Chutes
   #  Start
        16  => {end =>   6, action => 'down chute'},
        48  => {end =>  26, action => 'down chute'},
        49  => {end =>  11, action => 'down chute'},
        56  => {end =>  53, action => 'down chute'},
        62  => {end =>  19, action => 'down chute'},
        64  => {end =>  60, action => 'down chute'},
        87  => {end =>  24, action => 'down chute'},
        93  => {end =>  73, action => 'down chute'},
        95  => {end =>  75, action => 'down chute'},
        98  => {end =>  78, action => 'down chute'}
);

# Define dice probabilities
my @dice_index = (0..6);
my @dice = (0, 1/6, 1/6, 1/6, 1/6, 1/6, 1/6);

# Create transition matrix
my $transition_matrix = Math::Matrix->new(map { [ map { 0 } (1..$matrix_size) ] } (1..$matrix_size));
foreach my $state (1..$matrix_size) {
    foreach my $dice_roll (@dice_index) {
        my $destination = $state + $dice_roll;
        if (exists $transition{ $destination }) {
            $transition_matrix->[ $transition{ $destination }{end} - 1 ][ $state - 1 ] += $dice[$dice_roll];
            next;
        }
        $destination = $matrix_size if $destination > $matrix_size;
        $transition_matrix->[ $destination - 1 ][ $state - 1 ] += $dice[$dice_roll];
    }
}

# Set initial state
my $state_probabilities = Math::Matrix->new(map { [0] } (1..$matrix_size));
$state_probabilities->[0][0] = 1;

# Loop through turns
my $turn = 0;
while($turn < 100) {
    $turn++;
    $state_probabilities = $transition_matrix->multiply( $state_probabilities );
    my $p_ending = $state_probabilities->[$matrix_size - 1][0];
    print "Game has $p_ending chance of ending after $turn turns.\n";
}

# Print results
$state_probabilities->print;

1. Yay, I get to use my statistics minor!
2. And, the product of n matrices T is the transition probability matrix for moving from state i to state j in n moves.
3. You can modify this in the code.
4. Hat tip to toolic on PerlMonks for the transition table code.
5. Fortunately for parents, this has only seemingly happened.

No related posts.

Related posts:

1. Liturgical Wordpress, part deux Speaking o
2. RefTagger Plugin for Bible References in Wordpress Logos, the
3. WP Greet Box is Great I just ins

Related posts:

1. Family Photos, May 2009 I’ve
2. Photos of Family Visits We had two
3. Easter 2009 Jennifer,

No related posts.

I did consider an alternative, Referrer Detector. WP Greet Box appears to be more friendly to my caching plugin, and is Javascript/AJAX based so that it doesn’t slow down page generation times. Referrer Detector also appears to have a few more rough edges just based on the comments and changelog information I was able to find.

Usually, though, I want to choose the plugin that has the biggest community because that plugin will be more likely to be around for the long haul – and on this account, WP Greet Box wins handily with nearly 40,000 downloads. Referrer Detector’s 5,000 was nothing to sneeze at, but the community’s vote seems to be squarely with WP Greet Box. So far, I’ve been very happy with it.

Related posts:

1. Liturgical Wordpress, part deux Speaking o
2. RefTagger Plugin for Bible References in Wordpress Logos, the
3. Just installed Wordbook Just insta

Radical transparency is one of the highest ideals of Freedom Software, even more so than democracy. In fact, many such software projects are not democratic at all; witness the numerous projects that have a “Benevolent Dictator For Life.” The ability to freely acquire, modify, and distribute software is, in essence, the ability to work with your computer as transparently as possible. Freedom Software projects also exhibit radical transparency in their governance. Often times, even if decisions aren’t being made democratically, it is possible to know exactly with which data and why they were made. This sort of transparency is a natural result of software engineering tools. Bug trackers, mailing lists, source code repositories, blogs, wikis, Internet Relay Chat, etc. are all geared towards an egalitarianism of ideas that manifests itself in radical transparency.

Radical Opacity

Churches, on the other hand, are usually radically opaque. Like Freedom Software projects, they are not often run democratically, but rather have a Benevolent Dictator for Life (the pastor) or an oligarchy of elites (the Staff and/or the Church Council) who make all of the important decisions. But because churches have not embraced the transparent technologies I listed above, data for decision making remains locked up where the congregation and community can’t see it. Occasionally, there are good reasons for this – confidentiality being one. But more often, arrogance or ignorance become the stumbling blocks for wider distribution of decision making information.

It’s not a new idea to improve communication between decision makers and the rest of the congregation. The symptoms of this poor communication are not new either: people feel disenfranchised, don’t know what is going on, or are angry at decisions made without their knowledge. Poor communication is not a technological problem, it is a social problem – that is, a problem of the church’s leadership culture. Yet failing to take advantage of the opportunities afforded by the ethos of radical transparency is foolish. The technologies that enable Freedom Software projects to collaborate asynchronously and with less stratification can be a solution to opacity in the modern Church.

Ideas for a Radically Transparent Church

Consult your congregation’s computer geek for ways to get some of these started:

1. Set up a church email mailing list. This can enable discussion between members and get members’ opinions informally.
2. Set up a church wiki. Let members edit it freely to share information about church activities.
3. Blog everything. Every sermon, every newsletter item, every church council report, every treasurer’s report.
4. Video church council meetings and post them online. If something sensitive needs to be discussed, move to go into “executive session” and explain the reasons why before turning off the camera.
5. Have a church bug tracker. Let it become a place where the church’s material needs are tracked. This includes both little things (dripping faucet in the men’s room) and big things (family lost everything in a fire and needs new clothes).

Radical transparency doesn’t have to be technological. Nothing beats face-to-face conversation. Here are some things that can be done without a computer:

1. Coffee with the council. Once a month, set aside coffee hour to be an informal come-and-go panel discussion with the church council. Let the members talk – the council should mostly listen. Or spread the council members out, one per table, so that discussion is even more personal.
2. Regular pastoral visits at home. The practice of pastoral visitation of people who aren’t elderly or sick has fallen out of practice in many churches. Visiting with people one-on-one in their homes is a good way to understand your members, what they think, and what they want from church. It’s also a tangible way to bring the Gospel into the other 167 hours every week.
3. Post everything. A copy of (nearly) every scrap of paper distributed at every committee or council meeting ought to be posted prominently on a bulletin board in the church building.

What other ideas do you have for making more radically transparent churches?

Related posts:

1. Starting Point for Freedom Software in the Church As I write
2. The Coming Death or Rebirth of Church Publishing Every chur
3. “Freedom Software” versus “Free Software” “

You’ve got to know when to hold ‘em, know when to fold ‘em… – Kenny Rogers

I’m shutting down WartVid. WartVid was a project to get the Wartburg Seminary community to save money on DVD rentals and Netflix subscriptions by listing each person’s collection online and making it easy to share. It had a lot of great accomplishments:

• Over 1600 DVDs available, across a wide number of genres
• More than 60 enrolled participants, representing nearly a third of Wartburg’s on-campus students and faculty
• Written in fewer than 10 hours, total, by using the excellent Perl MVC framework Catalyst

Yet after the initial burst of enthusiasm, interest waned. I found that I didn’t even use it much myself, which is a bad sign for a hobby project. When the site starting turning up Internal Server Errors a few months ago, nobody even noticed, let alone complained.

So long, WartVid. It was fun, while it lasted.

Related posts:

1. Host your own OpenID with phpMyID Just got d

Well, a little Googling brought me to UUDeview, a program which hasn’t received any updates since 2004, but nonetheless is highly useful and available in your friendly neighborhood Ubuntu repository. Getting the images out was simple.

1. sudo aptitude install uudeview
2. Download the “original source” of the email to a file, let’s say /home/ted/Desktop/email.eml
3. uudeview /home/ted/Desktop/email.eml
4. Press p to set the folder where you want the images to go.
5. Press a to extract all the images.

And for those of you Windows or Mac users, it looks like there are versions available for those operating systems too. I’d say that this is the best little utility I’d never heard of.

Related posts:

1. Biblical Hebrew in Ubuntu Hardy Heron ere’

No related posts.