Maybe NASA should try Agile
I just read a short article, Space Sticker Shock in Scientific American about several projects at NASA having greater than factor of two cost overruns.
- The Mars Science Laboratory starting estimate .65 billion, 2 billion and going
- James Webb Telescope starting estimate 1 billion, 4.5 billion and going
In the article they mention a very interesting fact, that NASA no longer has labs and standalone projects for people to experiment in so that they have no off the shelf ideas or places to go to. They like so many companies are in the waterfall world, where plans get created and then get thrown over the wall for engineers and assembler to execute on. How many software projects have you seen stretch from months to years and then you see another team that does it in a few months with a totally different management approach. It is my belief that those types of teams should be blown up and reorganized, it is cheaper than using what they have done so far. These types of plans are great when everything is understood and you don’t need or want innovation. Clearly designing some of these device like software systems has many unknowns and complex interactions that no plan can account for. An Agile approach to building a Space product may reduce costs and produce better products or at least products where the over run will be detected early. Funding agencies should expect to have vaguer ideas of how much something will cost and instead change to an approach where you definitely get something that works but possibly terminate the program because it is too much to go further. One of the problems with Agile is estimating the total cost, perhaps that is the wrong problem to solve. Think of it as wanting a house that had 100 features but you only get 30, and the most important ones safety from the elements is non negotiable.
I really don’t think this will work for NASA, Agile hasn’t really proven itself in large projects (and these are mega projects) with huge teams. Not only that, I think that the instances of Agile implementation outside software development are very rare.
I have recently published an article about the limitations of agile, the article has a point for this particular issue (Point #3: Small Teams).
and one of these is from the Apollo mission where they had 2 daily standups to synch with each other and they had half day iterations where they committed to each other to have specific things done.
Maybe the problem is they GOT AWAY from being AGILE and decided that they had to act like professionals. the question is what kind of profession did they turn into. engineers, administrators or pencil sharpeners with a degree in media spin?
I had no idea that NASA was doing that in 1960. Thanks for the information.
I know many that can’t deal with twice a week meetings — consider that. Too many developers and engineers think the answer is to work totally independently IMO. With IM and microblogging I lean on all my friends and I try to help them. When you have a team that works this way there is nothing like it. Email and Forums just don’t work for many reasons. I run into issues and bugs in many products each day, no way I am going to login to each one and register and then wait for someone to flame me etc.