Apparently, at one point there was some comment that humans produced more CO2 than cars, which was supposed to mean that CO2 could not possibly be a greenhouse gas, by the classical naturalistic fallacy. It’s fun to debunk people within their own world views. And by fun I mean potentially more effective.

The average adult (resting) respiratory rate is 10-14 breaths per minute. [1]
The average adult tidal volume is 0.50 liters, with a total volume of 4-6 liters. [2]
Expired air contains 3.6% CO2 by volume, compared to 0.03% in atmospheric air. [3]
Each breath thus exhales 0.018L of CO2 ((3.6%-0.03%)*(0.50L)). Assuming STP, the weight of CO2 added by each breath is 0.035g. Doing the math, that means each human expires a minimum of 222kg CO2/year (489 lb/year).

The average car has a fuel economy of 17.2mpg. Each gallon of gas produces 8.79kg of CO2. [4]
In the US, the average per-vehicle yearly mileage was 12,016 miles in 2006. [5]
Doing the math, that means every car expires an average of 6130 kg CO2/year (13,500 lb/year).

The average car produces 27.6 times as much CO2 as does a baseline adult human. Furthermore, the carbon for human-sourced CO2 is from crops grown recently. However, the carbon for vehicle-sourced CO2 is from fossil fuels sequestered many millions of years ago. (Or, like most people making the argument believe, 5000 years ago when Noah built a boat to hold a breeding pair of millions of different species during a global flood).

But perhaps, you might say, there’s more humans than cars, so humans really are pumping out more CO2? Okay, let’s test that: There were, in 2002, approximately 590 million cars in the world. That’s one car for every ten people. [6] So if ten people expired more CO2 than one car, then breathing would indeed be more responsible for global CO2 than the cars. But a car expires the CO2 of 27.6 people, meaning that cars produce overall 2.7 times the amount that humans do.

I built a spreadsheet that lists the different powers that each player has, collects data about their attack bonuses and average damage, and determines the average damage the overall party can output in any given round. By design this is somewhat limited: Burst attacks, ongoing damage, and environmental tricks are all going to increase the party’s lethality, so keep that in mind.

At the end, I can essentially use the values from the spreadsheet as the basic monster defenses, and know that combat will take no more than the chosen number of rounds. And now that I have a hard limit on how long combat will take, it’s a simple matter to calibrate the monster’s damaging abilities to put the party in danger without being an outright party kill.

And so, without further explanation, the spreadsheet: MonsterHelper (Office 2010) and MonsterHelper (OpenOffice).

I’ve started working in NetBeans, purely for the extremely functional RAD GUI toolset. Hand-writing interface code in Java seems like an exercise in the purest frustration, and I’ve resolved not to gripe at Java software with crummy UIs anymore. I now feel your pain.

The code is in Trac at http://tepidpond.com/trac, and I welcome any submissions or suggestions anyone might offer under a GPLv2-compatible license.

So, I’d like a better replacement, something cross-platform and svelte. Imagine my surprise when I went looking on the Wizards of the Coast website for licensing information, and discovered that all third-party software is prohibited under their license! I suppose I’m a little late to the party; the fourth edition of D&D has been out for quite a while. But I’m offended. This is not how you make a geek happy, Wizards.

So, I’m now not building a replacement in Java. It won’t be released as open source. It won’t replace their shoddy excuse for a character creator. And it won’t act to encourage purchase of official WotC products. I am, however, going to build an Avatar Designer that will be compatible with certain game systems that are not subject to an abusive licensing scheme.

Watch this space.

1. Keep the Game Moving
Remember that you and your players are there to have fun. Getting hung up on something trivial in the rules is going to destroy immersion, cause boredom, and can result in players wandering away from the metaphorical table.
If a rules question arises, and you don’t immediately know the answer, invent a ruling. Usually there’ll be two or three possible interpretations of something a player wants to do. Pick the one that makes the most sense, moves the story forward, or makes the party happy. But it matters much less which ruling you deliver than that you deliver any ruling. Keep the game moving.

2. Prepare, but not Too Much
A group of experienced role players can usually tell which way your railroad is leading. But even the most seasoned level 10 cleric will surprise you, and eventually you will find your party in a situation that you haven’t written up dialogue for. And that’s okay! After all, the whole idea of roleplay, apart from the purely hack-and-slash fun, is to collaboratively tell a story. Maybe you’re interested in telling the story of the five heroic gnomes who save a mine from a rust monster, but your players are more interested in the ‘filler’ NPC you left along the road to give them directions.

Here’s the problem: If your filler NPC is no more than a line of dialogue, then when the rogue decides to investigate his backstory or pick his pockets you’ll need to invent the results on the spot. Right away, the effort spent to prepare the swarthy mine captain—a dwarf named Harinu Hammerfist—is wasted. If, instead of preparing the railroaded encounters with a laser’s focus, you decide to simply produce an entire world for the players to romp in, that problem doesn’t exist entirely.

But creating an entire world is hard, so don’t do it. Cheat! There should be a few characters that matter, with distinct personalities and stories. For the rest of the NPCs the party meets, you can simply tweak and reuse one of a few templates. So, invent a half dozen different NPCs, including at least a brief story and personality, and then insert them where appropriate, improvising dialogue where necessary.

You can do the same for battles, but more caution is needed here. If your sorceror dies because of a poorly-tuned random encounter, shortly before completing an epic Quest, her player would be readily forgiven for getting cranky.

3. Don’t Be Afraid of Murder
At some level, every player expects to die. Some times, they even look forward to it. But in a campaign where nobody dies, your party can start to believe they are invincible, and start to treat their world like it’s beneath them. Nothing ruins the suspense or intrigue of a story worse than a party member who rightly believes they are only an extended rest away from being out of any trouble. If your battles are continually too easy, or if your players are simply too powerful, they can lose their involvement and interest. If you notice your players getting this attitude, kill one of them.

But don’t settle for the cliché and have their death come like a boulder out of the clear sky. Weave their death into the story: If the leader of a cult decides they need a sacrifice, and your warden happens to be the perfect candidate, you could have him abducted from camp at night, dragged off to a cave, and ceremoniously offed just as his would-be rescuers crash through the door.

4. Give ‘em an Out
Don’t do what I just said. At least, not without privately consulting with that player. Roleplayers like to know they can trust cause and effect, and that if they don’t screw up they’ll be alright. If you kill a player’s character without any chance of a resurrection or reprieve, and that player isn’t in on the joke, you’ve likely just caused strife that’s going to ruin at least his evening. If you put your party into a situation they can’t win, or survive, or escape, they’re probably not having fun.

5. When All Else Fails, Summon the Kraken
Dungeons and Dragons players expect epic adventures and sprawling encounters. Why not give it to them? Involve them in the sweep of a country-wide conspiracy, let them take on a god, kill a great elder dragon, and shake the planet. We can all go wander around an inn talking mysteriously any day in the real world, and sparring with a foam sword is just as much fun against a friend as against a kobold. But Dungeons and Dragons is all about doing five impossible things before breakfast.

Now get out there and make some poor sod roll for initiative against a Kraken.

(II) XINPUT: Adding extended input device "G15 Extra Keys" (type: KEYBOARD)
(**) Option "xkb_rules" "evdev"
(**) Option "xkb_model" "logitech_g15"
(**) Option "xkb_layout" "us"
(II) XKB: generating xkmfile /var/lib/xkb/server-D378AD8F86E560F712A83EE36E4E5E92C595B9BD.xkm
G15 Extra Keys: dropping event due to full queue!
G15 Extra Keys: dropping event due to full queue!
G15 Extra Keys: dropping event due to full queue!

No output is shown in xev, and the inability to use my keyboard’s full range of powers makes me a sad panda. Sadly, this state of affairs has apparently been in effect since early April, as evidenced by the bug report on Ubuntu’s issue tracker. Oddly, that very same tracker has a user-submitted patch (dating from late April) which has been verified to fix the issue, someone reporting success with it, and…the very first developer response was two weeks ago.

But Ubuntu development timeliness is hardly my bailiwick. Here’s how to get the fixit patch if you happen to be an Ubuntu user with a G15, instructions shamelessly stolen from Niels Slot:

apt-get source g15daemon
cd g15daemon-1.9.5.3/
wget http://launchpadlibrarian.net/46395768/make-g-keys-work-again.patch
patch -p1 < make-g-keys-work-again.patch
dpkg-buildpackage
cd ..
dpkg -i *.deb

So there you have it, now your G15 works too.

AllowTcpForwarding yes
GatewayPorts yes

On home, add the following lines to ~/.ssh/config:

Host server
User username
IdentityFile ~/.ssh/id_dsa
ExitOnForwardFailure yes
RemoteForward 2200 localhost:22

On home, run the following command:

ssh -fnN server

This command must be running any time you want to make a connection to home.

Now, from any computer with a net connection, you can connect directly to home by opening an ssh connection to server on port 2200.

ssh server -p 2200

I think that this is becoming a measure in how bad Repubs screw up:
“Jenkins! What’s the fallout for me from Political Scandal X?”
“Sir, Fox is calling you a Democrat.”
“That bad?”

I mean what I said, though. Open Source is, or rather is infected with, a disease. The disease is a very simple one: “It works for me, so fix it yourself, noob.”

For some time, I’ve been using MythTV as my only TV system. For the most part, it works very well; recordings record, much TV is watched, and familial happiness reigns. But sometimes, and only sometimes, MythTV will stop working. Why does it stop? I don’t know. I look in the logs and find a segfault buried deep in a child process of the Myth frontend. Why did that make the whole system crash? I don’t know. If I was to be a good citizen and report this error, I would be told to go away. I would be told, if I really wanted to help report the bug, I need to install the debugging-enabled Myth, and sit at the keyboard with bated breath awaiting the crash. Then, I could type a sonata in G-minor into gdb(1), and get a proper set of debugging information. Finally, I would re-report the error, where a series of developers would debate the validity of a segfault in such a minor process, snidely blame my hardware, and eventually close the bug unfixed. Instead of diving into this deeply unsatisfying process, I simply told Mrs. Tepid to push the reset button frequently.

Why does Ubuntu make it so hard to play an MP3? If I plug a thumb drive containing a collection of 80′s-era German industrial rock into the slot of Windows, media player is there and can play it right away. If I find a Macbook and plug the same thumb drive in, it litters my drive with “.DS_Cache” files, but iTunes will play those MP3s immediately (although Genius may laugh at me for not being metro enough to use a Mac). If I plug my industrial ear-poison into a fresh Ubuntu install, Rhythmbox will happily try to open those MP3s, and cheerfully inform me that I need to download a codec. Of course I’ll indicate my consent to this codec business, but the opaque error message will leave me wanting. “Could not install codec”. So I dutifully search the internet for a solution, and soon come across this gem:

If an MP3 option is not shown in the Preferred Format list, install the gstreamer0.10-plugins-ugly-multiverse package and restart Rhythmbox. The option CD Quality, MP3 (.mp3 type) should appear in the Preferred Format list.

If, as I’m led to believe, evil media conglomerates prevent shipping the software with MP3 support built in, then why can there not be a button labeled “Install MP3 Support” somewhere readily visible? And by the by, you can repeat the entire MP3 rant about DVDs, as the crew of the shuttle Atlantis discovered.

I believe I know the root cause of the “fix it yourself” disease. It is simply this: open source hackers do not value time. With the exception of sponsored kernel developers, Mozilla employees, and a small number of other corporate-backed open source projects, these developers do not think in terms of time—whether yours or theirs—as being an asset.

It is this disease, and not any monopolist tendencies from other OS vendors, that keeps FOSS in the extreme minority. If it takes you three hours to setup a DVD player, that’s not a negative. If getting an OS to the state where MP3 files will actually play as music takes downloading massive codec packs with cryptic names, that’s all big media’s fault. It couldn’t be done by an automatic prompt, that would be catering to the RIAA somehow. And If making an iPod talk to a computer requires a haiku on the command line, that merely demonstrates the system’s power!

I use all three of the major operating systems (BSD, Linux, Windows). By far the one with the most problems, the one that takes the most effort to resolve errors, is the Linux machine. BSD just doesn’t break, and Windows has a nifty roll-back that’s fixed everything so far. But between dependency hell, recompiling, scripts that assume I’m on RedHat 3, wifi that still refuses to even give an error message before failing, and crashing video drivers, Linux is a large time sink with very little return.

Rant over.

I’ve had something of an epiphany. Vibrating motors might not be the best mechanism, if I want to maximize runtime and minimize noise. Better, perhaps, would be a series of inflatable air-bladders. The air would most simply be provided by a compressed cartridge. To direct the air, I’d want one solenoid valve per bladder for inlet.

As density of CO2 is 1.9769g/L at STP, a 12g CO2 cartridge contains about 6L. They cost about $0.65 in bulk. It should also be possible to use larger or refillable cartridges to improve running time.
Latex bladders are relatively simple to fabricate at home, acc to eHow.
Electronic solenoid valves would be used to direct the air to a specific bladder. Mead has a series that look suitable.

The general idea would be to use four bladders: one each for belly, back, left and right sides. The firmware would wake up once a second (might need to tune the interval) and inject the bladder(s) facing north with a quantity of CO2. The bladder(s) would gradually leak, likely remaining only 1/2-1/3 full in time for the next air charge. If they were completely sealed, an additional set of valves would be required to empty them; this would increase battery load and equipment cost, but conserve CO2.

Assuming the bladder would contain 1 cubic centimeter of CO2, and one bladder was filled per cycle, a single cartridge would provide 6,000 cycles, or about 1h40m. That’s not going to work. I want at least twelve hours out of a single charge, so I’d need about 90g of CO2. I found a source for 88g CO2 cartridges for $6.50 each, no idea of scale. How big are those?  They’re expensive, but in the right area for capacity. Also found refillable tanks rated for 4500 PSI that might work, but those are definitely too big: they’d add “backpack” to the list of hardware for the belt.

I think CO2 or HPA might be too much work to be worthwhile. I’ll look into pumps later. For now, chore time.

Thoughts?