Mild-Mannered Canadian Fury

Doug Stephen is Politely Peeved


Tue, 02 Jul 2013 «permalink»

The culimation of the VRC has resulted in me doing two things: sleeping a lot, and helping the lab manage press stuff. Now that things are a little quiet, I’ll be taking a vacation; I’ll be in Chicago tomorrow, then I leave for London on the 4th. I’ll be in London on the 5th and 6th, Amsterdam on the 7th and Delft, The Netherlands on the 8th, Berlin the 9th - 11th, and Prague the 12th - 14th.

In other words, you won’t be hearing much from me the next two weeks. Which shouldn’t be all that different from how it’s been most of the summer.

Stay cool.

IHMC Finishes First in International Robotics Challenge

Thu, 27 Jun 2013 «permalink»


Pensacola, FL – A team from the Florida Institute for Human and Machine Cognition (IHMC) took first place in the initial stage of the DARPA Robotics Challenge, topping a roster of 26 of the top robotics research groups in the world.

IHMC’s team scored 52 out of a possible 60 points in the DARPA Virtual Robotics Challenge, a computer simulation using software that will power a real-life humanoid robot in the future. Members of the top nine teams in the scoring will move on to the next competition, with the top six getting funding from DARPA (Defense Advanced Research Projects Agency) and provision of a humanoid robot, built by Boston Dynamics, for the next stage of the two-year competition, which is almost halfway completed.

The next competition, using actual robots, is scheduled for December 2013. The final challenge is set for December 2014, with $2 million in prize money at stake.

DARPA’s goal is to develop technology allowing the use of advanced humanoid robots at disaster sites, such as the Fukushima nuclear power plant in Japan in 2011, where human action is limited. The robots must be able to work in a human-oriented environment, and be supervised by people with little training in robotics.

According to DARPA, existing robots are highly specialized and very limited in their abilities; success “would mark a significant leap forward for the field of robotics.”

In its showing, the IHMC team finished ahead of a field of teams from universities, government agencies and private companies in the United States and overseas. U.S. competitors included teams from the Massachusetts Institute of Technology, Lockheed Martin, NASA’s Jet Propulsion Lab and Caltech, as well as teams from Spain, Japan, Britain, Mexico, Israel, Brazil and Poland.

IHMC fielded a 22-member team led by Research Scientist Jerry Pratt and Research Associate Matt Johnson.

“This is a big win for IHMC,” said institute co-founder and CEO Ken Ford. “Our team was up against some of the best and brightest robotics scientists and institutions in the world, and came out on top. We’re very proud of the team.

“This sort of recognition can only be earned by performing at a very high level, and that’s what our team did.” The Virtual Robotics Challenge required controlling a humanoid robot in a computer simulation to complete three tasks that could face a robot at a disaster site:

1) Get into a small vehicle, drive it to a specific location, and get out;

2) Walk across progressively more difficult terrain, including flat ground, a knee-deep mud pit, low hills and a debris field littered with cinder blocks;

3) Pick up a hose from a table, connect it to a spigot, and turn the valve on.

The competition was spread over three days earlier this month. The teams did 5 simulation “runs” on each task, with slight changes in the details, such as the height of hills or how the debris was spread out.

Multiple IHMC team members participated in the various challenges. For instance, Dr. Pratt said, 12 different people helped control the robot during the various trials. “It took four of us to drive the car,” he said. Problems had to be fixed before the next run, a reality that led to “all-nighters” to rewrite software.

Despite some early problems—the team’s robot fell several times, and had trouble at first driving the vehicle—IHMC’s team quickly mastered the challenges. The researchers had come into the competition with particular expertise in bipedal walking, said Dr. Pratt.

“Walking has been our strength, but we have learned a lot,” he said. “We did really well on the hose task. And we do well at human-machine interfaces. Any time we add autonomy to the robot, we make sure the human can interact in a positive way. The human knows what the robot is doing, why, and how you can modify what it is doing.”

IHMC is scheduled to receive the actual robot—named Atlas—in July for development work leading to the next competition in December.

“This stage was all simulation,” Dr. Pratt said. “We’re ready to use a real robot. We’re looking forward to the challenge.”

To read the release from DARPA, visit

For more information on IHMC, visit or contact (850) 202-4400.

The Institute for Human & Machine Cognition (IHMC) is one of the nation’s premier research organizations with world-class scientists and engineers investigating a broad range of topics related to building technological systems aimed at amplifying and extending human cognitive, perceptual, and physical capacities. IHMC headquarters are in Pensacola, Florida and a branch research facility in Ocala, Florida.

Members of Top Nine Software Teams Move Forward from DARPA’s Virtual Robotics Challenge

Thu, 27 Jun 2013 «permalink»

It’s been so hard to not write about this since the competition closed, but now that the press embargo is lifted I’m happy to share with you that the Institute for Human and Machine Cognition (IHMC) has finished in a strong, resounding 1st place in the DARPA Robotics Challenge (DRC) Virtual Robotics Challenge event (VRC).

It’s been an incredibly fun and challenging path getting here, and the competition was unbelievably fierce; going up against juggernauts like CMU, MIT, Lockheed, Case Western, and all of the other exceptionally brilliant robotics groups that competed was daunting. There were many, many days that we felt like we didn’t even stand a chance, so coming out with a showing like this for us has been fantastic.

All of the final scores for the top placing teams can be found at the source link. Thanks for putting up with the low post volume here on the site while I worked with the rest of my incredibly talented friends and coworkers at IHMC to make this a reality.

From the bottom of my heart,


DARPA and OSRF Robotics Challenge AMA Tomorrow at 1PM EST

Wed, 26 Jun 2013 «permalink»

Gill Pratt, the Program Manager for the DARPA Robotics Challenge, and Nate Koenig, the CTO of the Open Source Robotics Foundation, will be conducting an AMA on Reddit tomorrow at 1PM EST.

I always suggest tuning in to these sorts of DARPA events, because they usually coincide with some sort of press announcement or discussion of the events in question. The AMA is focusing on the DARPA Robotics Challenge (or DRC), spiritual successor to the Urban Challenge and the original Grand Challenge — the competition that spawned Google’s Self-Driving Car initiative. The DRC is what a large group of us back at the lab have been working on since last October when it was announced, and it’s been a wild ride. We’ve had our share of ups and downs, and it’s still early days in the contest.

Gill and Nate will most definitely have some cool stuff to share with us tomorrow. Stop in on your lunch break.

Differences Between Xcode Project Templates for iOS Apps

Sat, 15 Jun 2013 «permalink»

If you’re interested in iOS dev and aren’t reading Ole Begemann, you really should be. His post volume is low, but the signal to noise ratio is just so fantastic. XCode has a lot of layers and complexity tied in to its tools, including a dizzying array of templates, options, and configurability. One of the things that I struggled with when I started to teach myself iOS stuff waswhich template is right for me.

This is a great little primer to help sort that out.

And, if you’re a crotchety old-school UI guy like me that prefers skipping Interface Builder or Storyboarding1, you may be interested in Austin Louden’s 100% programmatic Hello World tutorial.

  1. One of my biggest pet peeves with NIBs: they’re a horrific pain in the ass to track under version control properly.

A Short Comment on PRISM Denials

Sat, 08 Jun 2013 «permalink»

I’m not really going to comment on the whole PRISM thing itself. I don’t have nearly enough information or expertise. Obviously, the claims being leveled are pretty atrocious, and if they’re even fractionally true then it makes me pretty sad about the state of American government. But what I do have some background in is the sensitivity of top-secret leaks.

Because that’s the one thing that nobody seems to be remembering when they report on Facebook and Google and other companies denying involvement with the NSA and their PRISM initiative. The program itself is classified. I worked for a small civilian defense contractor for a year and a half as a software developer, and it was during my time with this organization that the Wikileaks/Bradley Manning stuff really kicked off. And because of this, the government really buckled down on requiring their contractors to complete courses and training about sensitive material. What a lot of people don’t understand is that when classified material is leaked, it does not become declassified.

While the NSA may have declassified portions of PRISM, any information that has not been declasified explicitly is still considered legally protected in spite of anything that anybody has reported. This means that if any organization were to discuss anything about PRISM that is still classified, the government would have the ability to bring down a pretty brutal legal hammer.

This isn’t a treatise on morality or ethics or my stance on the subject. Personally, there’s a reason that I no longer work for defense contractors (which is where the money is in high-tech where I live). But the PRISM situation is more than just a PR nightmare for any company that is involved; it’s a vicious legal one as well. So take that in to consideration when you judge these corporate juggernauts (as shitty as their involvement would be if it’s true) when they protect their very existence by not committing treason.

Introducing PlexConnect, an AppleTV client which Thinks Different

Wed, 05 Jun 2013 «permalink»

Plex is an amazing home theater/media management solution that runs on almost every single piece of hardware that you can think of; it also uses a “server” model where one computer in your home can run a Plex media server and other devices can act as clients to stream content off the device.

There has never been an official Plex client for the Apple TV, though there has classically been a jailbreak app. This is the first time that Plex has been available on the current Apple TV’s without a jailbreak. Very excited to try this out.

A Chemex Iced Coffee Technique

Wed, 05 Jun 2013 «permalink»

It’s no secret that I love coffee. And for those of you that don’t live down here in Florida and have never been here this time of year, let me share a little secret with you: It’s hot as shit this time of year, and only gets hotter through August.

So it shouldn’t be surprising that a non-trivial portion of my adult life has been spent on trying to find the perfect iced coffee. It’s a finnicky drink to get right, with all sorts of caveats and workarounds.

Coffee itself is a complex extraction of oils, and oils do weird things when exposed to light, air, and heat. Hot brewed coffee can have it’s flavor changed really quickly by just sitting around, which is my most iced coffee methods involving hot coffee tend to revolve around making a sort of high-octane “concentrate” that goes directly over ice, getting diluted. On the other end of the spectrum we have cold-brewed coffee, which is a hit-or-miss topic with a lot of people; many of the flavors in coffee require heat to be extracted, so cold-brew just tastes different. Some people like it, some people don’t. I, personally, don’t mind cold-brewed coffee, but there is something to be said about the flavor of hot brew.

Up until recently, I had been using Marco Arment’s AeroPress Iced Coffee Recipe with great results. But in my never-ending quest for perfection, I decided to do a quick Google Something1 search for other ways to make iced coffee using the other gear in my armory. Specifically, I was interested in seeing what people had come up with for making iced coffee in a Chemex. To that end, I can’t take all of the credit for this idea, but I have made some tweaks to it. My roommate and I have been drinking this for a week or two now and it’s fantastic.

The idea is based on the third technique in this iced coffee article on OQ Coffee; the cold bath method. The idea is to submerge the Chemex carafe in an ice bath so that the coffee becomes instantly cooled as it drips down. This allows you to get all of the flavor of our hot brew without it staling/breaking down from sitting at higher temps for an extended period to cool down or diluting it with ice. My primary tweak to the recipe is that I don’t make a standard Chemex brew; I use a finer grind and a more dense coffee ratio so that I still get a sort of “concentrate” and a slower drip/better cooling effect. Even though the ice won’t melt as quickly, it’ll still melt pretty fast here in Florida. Plus, iced coffee is one of the only times that I do enjoy adding cream and sugar2 to my coffee.

With that out of the way, here’s my Chemex iced coffee “recipe”:

  1. Weigh out your water and coffee. My standard ratio for my Chemex is around 0.0625 grams of coffee per gram of water. For iced coffee, I like 0.09375 grams of coffee per gram of water. For example, my standard carafe of chemex is around 40 grams of coffee with 640 grams of water; an equivalent for iced coffee is 60 grams of coffee for 640 grams of water. Probably a good idea to pre-measure your water, since the weight of the ice bath might be a bit much to set the tare on a kitchen scale (I know that my cheapo scale can’t take that much weight).

  2. Bring your water to boil, and grind your cofee. I tend to go a little finer than normal; on my uncalibrated Baratza Encore, where I typically use an 18 for Chemex, I go down to around 12 or 13 for iced Chemex.

  3. Place the empty Chemex carafe with rinsed filter in a large empty bowl, and fill with a healthy amount of ice and a little bit of water. Heed the warning from the original site; a chemex full of air will get bobbed around if you use too much water, and could tip over or damage itself. It helps if you add the coffee to the filter before making the ice bath to add a little extra weight.

  4. Brew your chemex similar to how you normally would, blooming the coffee with a small amount of water for 30-40 seconds before finishing the pourover, avoiding the edges. Because the ground coffee is so much finer, I like to start off the first pour after the bloom by adding a little more water than I normally would and then using a bar spoon or similar tiny stir stick to agitate the water and grounds by stirring. After that, just keep the water level pretty constant until you’ve finished the pour.

  5. Enjoy. I like to portion out the coffee in to several mason jars, a 60/640 is good for about 3 or 4 drinks. I put the mason jars in the fridge, and then grab one, add some half-and-half + simple syrup, ice, lid, shake, and out the door in the morning.

  1. Hello darkness, my old friend.

  2. Or more properly, simple syrup if it’s iced coffee.