The last five percent
Posted Under: computer science, economics, management
When I was a software engineeer we had a saying: “The first ninty percent of a job takes the first ninty percent of the time and the next five percent takes the next ninty percent of the time…”
Of course the implication was that the last five percent was more or less never done.
When managing any project it is very important to understand when something is good enough. I found thed best way to do engineering projects is to do it in whatever way gets something working and slowly improve the project through stepwise evolution.
One of my favorite projects was when I programmed fifty transitions (dissolves and so on) for a non-linear video editing machine.
The original prototype code I was given to demonstrate the hardware capabilities was thirty kilobytes and was a simple dissolve.
My first version of the code did eight transitions. This allowed me to understand how the hardware worked. I next improved the software by evolving it to add more features.
The software was basically composed of a table driven graphics engine and a small table for each transition. Each time I would improve the software by making the engine better. Of course this wouls increase the parameters in the tables and I would just add the new parameters for the old transitions. In fact the way the engine developed was that the old tables would just end up with nulls (zeros added) for parameters needed for newer more complex transitions.
The final product ended with fifty transitions and took just slightly less than fifty kilobytes.
As I’m writing this I remember that one could say that anything done really creatively is channeled. That is to say one is most creative when one allows one’s unconscious mind take over and do the task. That’s how the above prject went. It was so easy.
I simply got out of the way and allowed the software to write itself like a beautiful painting.
***
Back to the last five percent. A project can never be done one hundred percent. The nearer a project hets to total completion (whatever that may mean – say fewer bugs discovered per year) the more difficult to get rid of other problems. That last five percent ends up costing many times more than the first ninety five percent which typically is plenty good enough.
In good project management it is important to know when good is good enough.
Winter Clearance Sale at VistaPrint! Save up to 90%
Post Footer automatically generated by Add Post Footer Plugin for wordpress.
