Monday, September 07, 2009


According to (pdf) NASA Manager's Handbook, [p.3-4, Table 3-4], the effort ratio between 10 years experience and 1 year experience is 2.6/0.5 = 5.2. It means that a single experienced individual is on the average 5 times as productive as a novice. Areas where experience is useful:

* Clarifying the problem definition, solving the right problem. During this phase you sometimes even see that there is actually no problem or that the problem is unsolvable. So you don’t waste time.

* Choosing the right approach/tools to solve the problem. Examples:
** For derivation of equations is the Newtonian or the Lagrangian approach more appropriate?
** Should we write the equations as vectors or should we use scalar components?
** Matlab vs C++
** Looking it up on the internet vs writing from scratch

* Knowing the right resources/people to summon/query

* Reusing you own documents/code written previously

* Getting things done due to personal power

* Minimizing blunders and information loss through disciplined data/knowledge management (source control, detailed documentation etc.)

* Being more productive with the choosen programming language.
** Knowing the language capabilities
** Knowing critical test cases; common pitfalls (floating point comparisons etc.)

* Being able to interpret analysis results, especially deviations from normal/expected.

No comments: