He’s making a very important free market argument here. The cathedral style of programming is very kin to a central planning scenario. The effectiveness and efficiency of any service, budget, product, etc. will always be only as good as the intelligence, creativity and desire of the few central planners that are assigned to the task. Converse this with the free market where thousands upon thousands of people have the chance to apply their skills and creativity to each problem in each industry. Distributed knowledge is impossible to replicate in a centrally planned model. But, more than that, it’s also impossible to replicate distributed creativity and even distributed “luck.” Raymond knows very well that so many times a computer problem is solved simply by pure luck. The right person being in the right place at the right time to catch that fleeting glimpse of what truly caused a given software bug. And that person is irreplaceable as a market participant.

This is just as true in the broader economy as it is in the software market. How many of our most essential products began as derivatives of another’s work and were perfected by the creativity of one guy in the right place at the right time? How many were discovered simply by accident? Many. Raymond understands this and is willing to take the heat for defending it. He’s been one of the few technologists who haven’t demonized capitalism, with such infamous quotes as “love doesn’t scale,” where he’s making the point that if we depend on getting critical goods and services from others simply on the basis of benevolence instead of a profit motive, we’re going to be up the creek. Love stops when the other guy can’t pay his bills any more. He’s also been critical of some moves within the open source community to police it’s own:

This is a perfect example of a good theory proving consistent down to a level of practice. He’s taking what he knows to be true in principle and putting feet to it in the real world. But, there aren’t a lot of his type in the open source community as a whole, which is strange. Most are an odd mix of some kind of government mandated “freedom.” Doc Searls is an example of this type of thinking. Searls generally agrees with Eric Raymond’s software development model on the whole. If he didn’t, he would hardly be the editor in cheif of Linux Journal magazine. His application of those ideas to a free and open market are disturbingly mingled with calls for government regulation, control and spending. He has called for government rollout of broadband services to citizens as an effort to create his idea of a consumer controlled economy:

But Doc misses the obvious in that you don’t own what we create. When a government uses the collective capital of it’s people to create a certain service or product, that product is the property of no-one in particular. The asset created is now the property of the collective whole, since the capital used was from the collective. This might sound utopian and wonderful, but in the real world it produces nothing but problems. The problems created by a property that has no invested owner are obvious. We love to praise the interstate highway system, but we conveniently leave out the horrible shape it devolves into before being fixed. We overlook the miles and miles of ghost construction that never seems to finish because no “owner” has a profit motive for completing it. Indeed, we can’t “begin to calculate what it would cost us today not to have it.” But that isn’t because the interstate system was such a wonderful idea. It’s because we don’t know for sure what would have taken it’s place. We might have something much, much better today if central planners hadn’t wasted all that confiscated capital on the limited vision of a few central planners. Futurama-style people tubes anyone?

The difference between Raymond and Searls is the dialectic. The one side respecting the connection between capital and private property, and seeing the unhampered expression of those as producing the sought after results by logical necessity. The other side is trying to force an open ideal through the heavy hand of government – producing a bazaar that is carefully constructed and regulated by a staff from the biggest “cathedral” of all. You might call them the “priests” of capital and risk. I’ve been glad to see Jeff Tucker and other’s from the Mises Institute so interested in open source and the idea of anti-intellectual property. It will be libertarian free-market types that will ultimately swing this argument in the open source community to the right conclusion. The conclusion that government meddling in capital markets produces not more openness, freedom and competition, but more fascism and corporate rigidity. And that is a detriment to all markets. Especially open source.

Critical listening on this subject:

Against Intellectual Monopoly: Michele Boldrin Interview:


Download

I went on the search for a cheaper alternative and found the 1-wire temp/humidity/flood sensors provided by iButtonLink. They are just the bare sensors with serial connectors for interfacing. The software to query the sensors is called digitemp and runs on Linux and Windows. We already have a monitoring script that runs every 10 minutes to check if certain critical services are down, so modifying it was easy. You can do it with some simple Bash code:

mailto="admin@domain.com"
maxtemp=75

temp=`digitemp -t0 -q -o3 | tr ’.’ ’ ’ | awk ’{print $2}’`
if [ $temp -gt $maxtemp ]; then
   echo "The server room is above the critical \
   temp threshold of $maxtemp." | \
   mail -s "CRITICAL TEMP: $temp" $mailto
fi

Just run that from a cron job every few minutes and you’re good to go. I’m not interested in humidity/water monitoring at the moment, but I’ll post an update as soon as I get those running with some more code.

# touch /root/mysql-nightly1.sql
# touch /root/mysql-nightly2.sql
# touch /root/mysql-nightly3.sql
# touch /root/mysql-nightly4.sql
# touch /root/mysql-nightly5.sql

Here is the script:

#!/bin/sh

##: Go home
cd /root

##: Keep five days of backups on hand
dumpfile=(`ls -1tr mysql-nightly?.sql`)
dbserver=’localhost’
dbadmin=’you@domain.com’

mysqldump --host=$dbserver --user=user --password=password \
 --all-databases > ${dumpfile[0]}
if [ $? -eq 0 ]; then
  echo "Successfully dumped MySql database from $dbserver."\
   | mail -s "INFO: MYSQL DATABASE DUMPED" $dbadmin
else
  echo "Error dumping MySql database from $dbserver." \
   | mail -s "ALERT: ERROR DUMPING MYSQL" $dbadmin
fi

##: Restart the database server
/sbin/service mysqld restart

cd -

Here is the cron job entry:

18 21 * * * /root/backup-database.sh > /dev/null

...
zone "foo.com"{
    type master;
    file "db.foo";
    notify yes;
    allow-update{
        localhost;
    };
};
zone "1.168.192.in-addr.arpa"{
    type master;
    file "db.192.168.1";
    allow-update{
        localhost;
    };
};
...

This gives processes running on localhost the ability to make ddns updates on those zones. Just change “localhost;” to whatever host dhcpd is bound to if it isn’t running on the same server. You can also specify multiple hosts or a whole subnet using the “192.168.1.0/24″ format. After you have configured BIND it’s time to tell dhcpd where to send updates. You do this by adding the following to your dhcpd.conf file:

ddns-update-style   interim;
...
subnet 192.168.1.0 netmask 255.255.255.0 {
  ...
  ddns-updates on;
  ddns-domainname "foo.com.";
  ddns-rev-domainname "in-addr.arpa.";
  ...
}
...
zone foo.com. {
    primary localhost;
}
zone 1.168.192.in-addr.arpa. {
    primary localhost;
}
...

This tells dhcpd where to send ddns update requests and for what domains. Be sure to monitor your syslog as you fire everything up to make sure you aren’t getting any ddns errors from either of the two daemons.