Confirmed or Busted: Are the Mythbusters Agile?
In 2009 IBM invited a couple of well-known TV personalities to appear as keynote speakers at an international event. They were interviewed by Scott Ambler, who many will recognize as the driving force behind initiatives such as agility@scale and Disciplined Agile Delivery. That event was the IBM Rational Software Conference, and the two guests were Adam Savage and Jamie Hyneman. The show for which they were famous? The Discovery Channel's Mythbusters.
This series is enormously popular with technical types, and it's fair to say that it has earned something of a hallowed position in nerd culture. It's certainly a common topic at the water cooler for many in the IT business. I reckon that most of you will have seen an episode at one time or another, and a significant proportion of you are likely to be addicts. I don't even own a TV, but I've caught the show in hotel rooms when I've been on the road and found it to be similarly compelling. So for the uninitiated, here's the low-down on what it's about. In each episode the team will set out to prove - or to debunk - certain myths and urban legends. For example, in one episode they tested the myth that if a bullet is fired horizontally while another is dropped vertically at the same time and from the same height, they would hit the ground simultaneously (result: myth confirmed). They followed that up by testing the idea that it is actually possible for someone to be "knocked out of their socks" (result: myth busted). With backgrounds in the special effects business, and most likely a nod to Wile E. Coyote, the Mythbusters use plumbing, joinery, metalwork, explosives, electric circuits and a battered crash test dummy to build Rube Goldberg contraptions that put oddball claims to the test.
So where is the connection between this "seat-of-the-pants" engineering and agile methods? Why is this show so popular amongst agile developers, even to the point that Adam and Jamie appeared at the IBM Rational Software Conference keynote with Scott Ambler? Is there something in their approach to busting myths that resonates with modern software development practice? Or to cut to the quick: is there any truth in the assumption that the Mythbusters are somehow "agile" themselves? That's the myth that I decided to investigate. Will it be confirmed or busted?
A Seat of the Pants Investigation
For this evaluation, I watched some re-runs that were showing on my hotel TV. I sat in front of it with my laptop, and with no clear idea of method, but before long I found myself typing notes about the things I was seeing. Here's what I came up with.
- First some initial impressions. Judging by appearances, this myth doesn't look too promising. Neither Adam nor Jamie look much like agile developers. At a stretch, backgrounds in a creative digital agency might possibly be indicated, but that would be about it. In all honesty I'd be more likely to peg them as a couple of artists. Adam Savage, the rangier of the two and with a light ginger beard, looks like Van Gogh before he went mad and cut off his ear. Jamie Hyneman, stockier and sporting a walrus moustache, has a trademark black beret that seems more Parisian than Palo Alto. As if to confirm his approbation of the arts and a disdain for the technological, Jamie has a university degree in Russian literature. It's easy to imagine these two as 19th Century Bohemian renegades, fleeing their garrets in the Sorbonne to set up a commune in the Camargue. Instead they run a special effects workshop in California, which I suppose must be the modern equivalent.
- Once you start looking beneath these bare appearances, the prospects for their agile credentials seem to get even worse. Due to the nature of the myths they test, IT infrastructure doesn't seem to feature much in any of these shows. The Mythbusters build things for real, in the old-school way, and not by computer simulation. The welders and greasers at your local body shop would seem to have more in common with them than software developers. This makes it hard to draw analogies with conventionally agile ways of working. Clearly it doesn't rule out synergies entirely - there's nothing that constrains agile practice to the software field or even to IT. It just means that it will be necessary to distil key agile principles, and determine the extent to which they might apply in this strange case.
- In the myth I'm now watching, the team are evaluating whether or not the doping compound used on the Hindenberg could have contributed to the airship's rapid burning. The idea is that the mixture could behave like thermite. They started by testing sheets of material with various ratios of compounds painted on them, including actual thermite as a control, before progressing to a replica model of the airship. I suppose they can be said to have used a test-driven approach. The initial focus on testing sheets of the material - rather than a reproduction of the airship - can perhaps be correlated to unit testing.
- Now they are seeing whether or not running zig-zag away from crocodiles and alligators is a good strategy for avoiding injury should one try to attack. The team built a contraption that swung a dummy forward in a zig-zag fashion. They tested it and redesigned the prototype before using it for real. Once deployed however, they have found the reptiles very difficult to provoke. They either seem to recognize that the dummy isn't genuinely a person, or they just don't feel threatened by it. The team have collaborated with each other closely on what to do to get the animals to co-operate, and although they have not been successful this myth has shown a good use of prototyping and I have seen numerous examples of teamwork, inspection and adaptation.
- The team are testing whether or not the ancient Chinese could detect an army tunnelling under the ground by listening against a submerged drum. The equipment they are using is fairly simple and by varying the position of the drum they can hear digging in a mineshaft deep below ground. There doesn't seem to be any evidence of agile practice in the evaluation of this myth, but in varying the location there is clearly some application of scientific method and the use of inspection and adaptation based on results.
- Now they are testing a suite of bacteriological myths: is food OK to eat if it falls on the ground for less than 5 seconds, is a dog's mouth cleaner than a human's, and is the toilet seat the cleanest place in the house. They've taken a number of swabs and cultured specimens on agar plates. A control plate was also used. Again, this suggests the application of scientific method but there is no conclusive evidence of agile practice. There may be tangential evidence in the use of a control, in so far as the "red-green-refactor" principle also uses negative testing as a point of reference.
- The team are testing whether a bullet that is fired horizontally from a gun will fall at the same rate as a bullet that is simply dropped from the same height, and will therefore hit the ground at the same time. They started off by using a pinball firing mechanism to shoot a bullet before progressing to a paintball gun and then to a real firearm. This is another test-driven approach that has made extensive use of functional prototypes. Adam and Jamie have discussed the results of each prototypical experiment before moving on to the next, and again the process they are following has shown clear (if irregular) inspect and adapt cycles.
- Now they are testing a myth about whether it is possible for someone to literally "be knocked out of their socks". It seems that at some point the Mythbusters had put out an appeal to fans to send in their socks so they could test a completely different myth about what happens to those that go missing in a wash cycle. However, by popular demand they have been asked to test this myth instead. If so, this may be comparable to a "pivot"...the Lean Startup technique of changing a value proposition to match the re-evaluation of market demand.
- This "Faked Moon Landing" myth is an interesting one, partly because of the controversial subject matter and partly because there is comparatively little evidence of agility in the team's approach. The team really went to town on the myth and have spared little expense, even to the point that simulated lunar gravity flights have been conducted in an aircraft. The experiments appear to have been stage-gated and have clearly been thought out carefully beforehand...though perhaps rightly so. Agility is based on inspect and adapt cycles, which to the naive mind can look like "making it up as you go along". Apparently there are millions of naive minds that believe the Apollo landings were faked. I can see that with a politically sensitive myth like this, you'd have to be careful to avoid any appearance of bumbling along, lest the field be taken by idiots who pretend to have surer answers. So perhaps the team are right to have adopted this comparatively prescriptive approach. It's not that the myth couldn't have been tackled in a more incremental manner, it's arguably more a case that under these circumstances it shouldn't be.
- The team are evaluating myths about the uses for duct tape. They've successfully built a duct tape cannon which fired a cannonball and stayed intact. A small-scale prototype was used first and only when this proved acceptable did they move to a full-scale model. They've also tested the idea of being able to build a boat out of duct tape. They debated how many layers of tape to use and made the decision to stop at two and not to bother adding a third, thereby potentially incurring technical debt for the sake of expediency. Adam and Jamie made bets on whether or not the craft would hold as a result of this decision. It did, confirming that their "definition of done" had been adequate and that the boat was fit for purpose.
So then, what does this brief examination of a handful of episodes say about the Mythbusters being agile? Does it take us any closer towards explaining their appeal to developers, and why they would be invited to appear as keynote speakers at a software conference? Before I started this cursory investigation I thought I could guess the answer. I thought the show was popular with developers not because the Mythbusters are "agile", but because uncool nerds want to associate themselves with cool dudes doing cool things, and which seemingly validate non-prescriptive, bare-to-the-metal practices. And to be honest I still think there's an element of truth in that. Yet the whole truth isn't so simple, and it seems to run deeper than I first suspected.
The show is pitched as a way of getting people interested in science, and the application of scientific method does appear to be a recurring theme. Agile practice is a reflex of scientific management and it too makes extensive use of observable results to facilitate the inspection and adaptation of a process. As such, the scientific approach taken by the Mythbusters cannot in itself be used as evidence of agility. There is also no evidence of regular time-boxing either for the elicitation of metrics or for the control of scope. The management of technical debt is erratic even though the principles and risks seem to be generally understood by the team. There is however very good evidence of team collaboration, and for the controlled inspection and adaptation of the myth testing process. In one case a "pivot" was done that echoes the market-driven agility of the Lean Startup movement (in their conference appearance Adam & Jamie mentioned rapid course changes by cast and crew when prompted by Scott Ambler to discuss their work in the context of agile practice). In most episodes there is a clear hacker ethic that underpins their methodology, and they rarely use prescriptive approaches to evaluate myths. The only exception that I saw was the debunking of the "faked moon landing" allegations, but this discrepancy may be explainable by context.
So are the Mythbusters agile? I'd really need to make a more detailed examination of the series to state an opinion. However from what I've seen so far, and as the team themselves would probably say...myth plausible.