A common question among project management staff is who should make the call to release a product, when they should make it, and on what basis they should judge. I can now definitively state I have uncovered the answer. I’ve been reading through historical records of past projects and have uncovered a system whereby countless projects from other disciplines were successfully launched to great acclaim. The system is ingenious in its simplicity in that all you need is a flock of birds and some one you can pay to tell you what it means when they fly in certain patterns. The Romans were adept enough at this practice to have kept detailed documentation on patterns, weather, calendar dates, and breeds and how each factor was interrelated into a complex system. They called these books “historical data” which they used to analyze their “metrics” before giving the blessing of the gods or not. The beauty of this system was twofold, first it guaranteed an income for the second and third sons of aristocartic families who otherwise might be forced to become farmers or merchants or something actually productive. Second it completely removed any responsibility for the eventual decision from anyone who might be punished for it or otherwise blamed later.
I apologize if this sounds like the rambling digressions of a bitter and possibly demented turnip, but the practice of augury and their practitioners (the augurs) have been much on my mind of late. I’ve also been thinking a lot about the ordeals of Hercules, particularly his task to clean the Stygian stables of manure. It was a seemingly hopeless task as the stables were so large that any effort to clean one area would mean the others ended up being fouled in the meantime. I know its not technical jargon, but we can’t all be computer science graduates now can we? Some of us actually wanted chat up girls in college, so we followed them into the Arts department.
My Stygian task of late has been to address some modern day augurs who seem to believe you can predict the point when software is to be released by studying the flight patterns and behaviors or software’s version of birds, or rather tests and defects. Every time I think I’ve addressed the subject with some wit and wisdom, I find the blessed group has snuck back into the stables with their wheelbarrows full of charts and studies ready to spread around with their loamy gooodness.
The ugly little truth about releasing software to the client is there simply is no objective metric by which you can judge. The emperor has no clothes. The birds fly that way because some one spilled grain on the street just over the wall. Or worse yet, they’ve been trained to fly that way because the augurs know you’ve done this enough times to have a few expectations. The fact of the matter is releasing software is about intuition. Informed intuition, but its still intuition. Some one or some group has to be tasked with “making the call” and they have to be given whatever information they need to make said call. But at some point, the call has to be made because not making a call is in and of itself making the call not to release. So-called objective metrics may be informative in the macroscopic realm of process improvement when analyzing multiple releases across multiple cycles and products, but they are somewhat useless predictors for specific instances. They can be among the information requested by the decision maker(s), but they should and cannot be considered the decision makers themselves.
Perhaps it would be better to switch metaphors and start talking about cricket or baseball. Both games appear to be ruled by statistics. Batters and pitchers/bowlers are swapped out because of a perceived advantage against a particular opponent based on statistical performance against similar persons. In both games the person batting must commit themselves to an action before the moment for that action (ie: the ball arriving) occurs due to the high rate of velocity at which the action is occurring. Both decisions, whether to swing and whom to play, are informed intuitive decisions based on statistical analysis and talent, but the analysis and talent cannot ensure the decision is correct. That’s why the games are (sort of) exciting (to some). Statistics can tell the batter the picther/bowler is likely to throw the ball a certain way, but they cannot assure him or her of that throw. The batter must commit to swinging before they know the actual outcome of the pitch/bowl, either connecting or not.
The same is true of software releases. Talent, experience, and statistics can help inform the decision that must be made, and any decision maker would be criminally foolish not to avail themselves of said resources. But passing the cup of responsibility onto the statistics themselves is similar to swinging your bat or not because statistics state the pitcher will now throw a curve ball into the lower left pocket of the batter’s box. Or to put it more appropriately, it would be like declaring war on the Etruscans because there were three white doves in the flock that flew south while thrumming. The war or the project will continue under its own momentum, but the decision makers charged with declaring war or release can absolve themselves of blame while reserving the option of showering themselves with glory. In reality the only ones who win are the augurs who charge for their services.
The rest of us just have to go about our business and clean up the droppings.