A Pattern language - SkunkWorkEngineering?
Rarely are SkunkWorks projects supported or encouraged by a company. It is a very difficult thing to manage and control from an official company policies point of view. Some corporations have dabbled in supporting SkunkWorks (Hewlett-Packard kept an 'open lab and supplies' policy to encourage tinkering...), but it is still rare. 3M had something like this. Do they still?
By realizing that you have a need to develop something in the SkunkWorks it does not suggest that a corporation is bad or restrictive. Change is hard to promote, and it must sometimes evolve from a grass-roots movement.
An example: You are in a corporation of several hundred people. The corporation as a whole will not make a move to Object Oriented Technology. Maybe the group you are in are reluctant to be the first to make such a move (lack of training, inertia, contract or customer constraints, etc). This is where SkunkWorks comes into play.
-- ToddCoram
Everyone is encouraged to modify, annotate and expand upon the following list of possible SkunkWorks Patterns:
- DesignTheRightThing - Work toward the right solution, even if a worse one has been suggested.
- SneakInNewTechnologies - Introduce new techniques silently.
- TwoDocsForThePriceOfOne? - If your documentation standard is lousy, supplement it with a more robust style (perhaps in Pattern form?)
- ProofInPudding - Make sure that you are successful, it is hard to argue against success.
- RevolutionThruEvolution? - Slowly change the way you develop software.
- HitHardAndFast? - Don't wait around for an audit. Take too long and someone will figure out that you are doing something different.
- AlwaysMeetDeadlines? - See LieToYourManager.
- LieToYourManager - As long as you deliver, don't mention the fact that you are trying new and different stuff... This can be dangerous though...
- RunSilentRunDeep? - SkunkWorks should NOT be high, medium or low profile, it should be sub-low profile.
- NoSilverBullet (If you meet the Buddha in the middle of the road, kill him) - Don't get trapped in the idea that you have found the perfect technique that will revolutionize the company.
- LordOfTheFlies - Don't soothe management, beat it to death.
- GiveThemWhatTheyNeed - If the ProofInPudding, and management doesn't know what to do because IcouldDoThisInaWeek? , then just GiveThemWhatTheyNeed.
- DontAskPermission - Sometimes it's a lot easier to just do the right thing than to get the appropriate signatures on a proposal to do the right thing.
PaulMcKenney sent me these additions:
- MakeApostlesSuccessful? - if the idea is really radical, you may need to prove that others can adopt it. Find some risk-tolerant colleagues and help them use your idea. Do whatever it takes to make sure that they don't fail.
- AlwaysHaveOneOnTheBackBurner? - Skunkworks projects are risky and prone to failure. The only way you will get a reasonable number of successes is to Try, Try, Again. This is particularly important if you are trying something truly new (as opposed to transferring proven technology from somewhere else).
- CultivateSenseOfHumor? -You will need it. ;-)
The Lockheed lawyers periodically remind people that Lockheed has a trademark on the phrase "Skunk Works". Such a reminder appears in a recent Doctor Dobbs Journal. -- DaveSmith
Insane. What are they going to do? Release a product with that name someday?
[Lockheed's tight claim to this trademark is particularly amusing when you
consider that skunkworks translates as our company's management would screw this project up if they knew it existed.]
Lockheed is jealous of the name. They got SCO to rename their "Skunkware" tools CD with legal pressure.
See: BasicOperatingRulesOfLockheedsSkunkworks
CategoryPattern