Differences between revisions 4 and 5
Revision 4 as of 2012-09-16 19:04:04
Size: 2107
Comment:
Revision 5 as of 2012-09-16 19:04:25
Size: 2109
Comment:
Deletions are marked like this. Additions are marked like this.
Line 11: Line 11:
 * Relevant: All experiments produce unexpected results, some are game-changers but most are merely interesting. But a mission-oriented research group or company isn't distracted by irrelevancies. Experiments that produce too much unusable data are no better than experiments that produce too little.  * Relevant: All experiments produce unexpected results, some are game-changers but most are merely "interesting". But a mission-oriented research group or company isn't distracted by irrelevancies. Experiments that produce too much unusable data are no better than experiments that produce too little.

"Experiments" with Balloons, Suborbitals, Cubesats, and LEO - Why not?

Many helpful people suggest Server Sky "experiments" using other technologies or other locations. This page discusses those suggestions, and why they may be an expensive distraction rather than a step towards orbital data centers.

Experimental Design

Proper experiments are designed to provide definitive answers to specific questions, generating as much relevant information with as little expense, effort, and delay as possible.

  • Specific: Experiments aren't shots in the dark. You start with a hypothesis or a question and test it. You want unambiguous results.
  • Generation: If you already know the answer, an experiment tells you little.
  • Relevant: All experiments produce unexpected results, some are game-changers but most are merely "interesting". But a mission-oriented research group or company isn't distracted by irrelevancies. Experiments that produce too much unusable data are no better than experiments that produce too little.
  • Expense: A cheap experiment is better, but nothing is more expensive than an experiment that doesn't provide definitive results.

Sometimes, an experiment is a test of a team, or a manufacturing process, or a prototype intended for production. However, there must still be specific questions to be answered in quantifiable ways. For example, we can model and predict the behavior of a team, then perform a task, then compare actual behavior to what we expected. Deviations call for improvements in our prediction process, and improvements in team staffing and management.

An experiment should always start with a hypothesis or a question. We are testing nature, but we are also testing the thought processes that led to the experiment. If we are vague in our goals, or settle for vague results, we've learned very little.

Balloon Radio Relay Systems

MORE LATER

Suborbital and Parabolic Flight Tests

MORE LATER

Cubesats

MORE LATER

LEO Testing - Free Flying and ISS

MORE LATER

BalloonsCubesatLEO (last edited 2013-02-17 02:31:44 by KeithLofstrom)