On each of our systems projects, building and deliverying remotely deployable systems that accept commands and delivery telemetry data, we use Perl to script all of our Baseline Functional Tests. We do system integration and testing over about 12 to 18 months; during which we test, extensively all commands and all of the telemetry points that the system can accept/provide.

However, we have rarely gone back and VALIDATED that we have done such an extensive composite of testing.

About a month ago, our management began to say "prove it to us." Prove to us that every single command and every single Telemetry point was tested.

We typically have about 800 to 1,000 commands and about 1,000 to 2,000 different telemetry points per system and all of our testing dutifully generates log files of everything that was done.

So I proposed that I could process the roughly 10,000 lines across about 800 log files to find the answer.

So I sat down and over about a 3 day period (most of which time was reviewing and being sure I knew the structure, syntax and oddities of the log filling processes) I developed and ran a fairly simple and uncomplicated script that generated exactly the results that were looked for.

Turned out that 1: The bosses were delighted and impressed and 2: we found that there were 3 commands that had not actually ever been tested...great BONUS!

I've tried similar tasks in the past with various languages (FORTRAN, VISUAL BASIC, C) and have always hated the process and difficulty of the task. Perl made it a true "no brainer"!