In the current phase of projects at work, a lot of people around me have been traveling, either from elsewhere to here, or from here to other places around the globe. Also a lot of others are taking summer vacation trips. With all this travel going on, one thing comes to mind: road rallies.
I was watching the 1956 Monte Carlo road rally and the narrative made me think of software testing. The narrative focused on the real purpose of the rally: car makers submit their best vehicles to an independent test team (drivers and judges) to test against an objective set of metrics.
The course is 2,600 miles across Europe. It takes about 72 hours if you include the add-on tests and awards conducted at the end of the 2,600 miles in Monte Carlo. Car makers had “teams” so there might be a six-person team in three Alfa Romeos participating in the rally. Likewise for Mercedes, Vanguard, Bristol, Lancia, Renault, Citroen, Jaguar, Austin Healey, Porsche, Riley, Sunbeam, and even two Fords, Anglia and Zodiac.
The purpose was to test every aspect of vehicle durability, maneuverability, speed, suspension, braking, engine integrity, to demonstrate overall automobile performance, endurance, and reliability. Cars that do well get kudos in the industry, and sell more cars. Car makers used the rally as a showcase to gain market share. Road rallies were marketing activities, driven by good test results.
During certain phases, no repairs are allowed, so you rely on well-made parts and cross your fingers. At one control point, you have to stop, then start down a steep hill keeping up a required speed through hairpin curves, then slam on the brakes and stop with the front tires between two white lines painted across the road. Outside the lines equals marks against you; and a mark against you for each one-fifth-second slower than the required time.
Throughout the rally, you have to drive through regular streets, not necessarily highways, up and down mountains in snow, meet the deadlines at control points along the way, but never violate local speed limits. Roadways were not blocked off for the rally, so in towns and villages, you mingle with local traffic. You can be penalized for going too fast as well as too slow. Navigators are constantly minding their arithmetic. Like software, the automobiles were judged by several different types of metrics.
When you think of road rallies, you think of a race. But the classic European road rallies of the 1950s and 1960s were a QC crucible of demanding repetitious tests. Judges took measurements at many “quality-control” points throughout the rally. European roadways became a test lab where the output was painfully public for car makers whose cars faltered, and pleasantly public for those whose cars pass the tests.
Like a true test lab, road rally awards come in different forms. Awards are given for safety results, a 150-mile steep-hill climb, the timed braking test mentioned earlier, as well as ultimately finishing closest to optimal time (not too early, not too late). All of these test areas have a point system. The overall winner has to have speed, performance, endurance, integrity, and fewest penalties.
Remind you of anything? Of course, software testing, which results are also measured by speed, performance, endurance, integrity, and fewest penalties. Maybe we should start a national test team for software—allow software to be submitted to a 72-hour QC crucible of a systematic beating and see who wins. Drivers and judges who know how to put a machine through its paces might help us improve software quality, which, as we’ve demonstrated many times, also means improving quality of life.
Also improving the quality of life, would be waking up in the morning and finding a 1956 Alfa Romeo in the driveway. Happy trails!