Two weeks ago distributed.net announced the beginning of its new project, OGR-27. This project will verify that the current shortest-known Golomb ruler with 27 marks is optimal (i.e. the shortest possible), or it will find the optimal 27-mark ruler. Bovine, a leader of the project, stated that “we are confident that we will discover a better ruler for OGR-27 than the one we know to be optimal currently.” He also stated that the project will take about seven years to complete.
So, how do you get excited about a project that will take so long to complete? Seven years is a long time. You may be in a different decade of your life seven years from now. You may be married (or divorced) seven years from now. You may have children (or your children may have grown up and left home) seven years from now. You may be in a different state (or province or country), a different job, even a different career. You will almost definitely finish the project on a different computer than the one on which you began it. In seven years you may not care about Golomb rulers, let alone finding the optimal 27-mark ruler. How do you commit to a project when its end is so far in the future?
Here are several reasons to get excited:
1. You know when the project will end. Because of the nature of Golomb rulers, there is no way to know exactly how many rulers there are for a given number of marks until you have found all of them. For its first OGR project, OGR-24, distributed.net had no way to estimate how much work needed to be done for the project so there was no way for participants to know how much of the project was complete or when it might finish. For OGR-25 the project owners discovered that while they couldn’t know the total number of rulers they would need to test, they did know the total number of “stubs,” or beginnings of rulers, they would need to test. Work units for each OGR project are individual stubs, such as 27/4-8-35-45-24* (which my computer is working on right now). They could show a project’s progress as the total number of stubs completed compared to the total number of stubs for the project. It’s harder to be excited about a project when you have no idea when it will complete. It’s easier to be excited when you know roughly when it will end, even if that end is several years away.
2. The project may end sooner. In the next four years most OGR-27 participants will have upgraded their computers and will have at least twice as much computing power as they do now. More participants may join the project as they see it move closer to completion. More people may become interested in Golomb rulers within the next seven years and may join the project. Some distributed computing teams may get into a stats competition for the project and may temporarily give it a boost in computing power. Assuming the project leaders did not factor these potential growths in computing power into their time estimate, the project could end in six or even five years.
3. You’re contributing to something new. No one has found the optimal 27-mark Golomb ruler before. You’re helping to make a new discovery, helping to make history. You’re also contributing to something big. Until a few years ago it would not have been practical to attempt a project of this size. Seven years is an easier length of time to commit to than 14, or 20.
4. Larger goals give you a greater sense of accomplishment. A runner can feel a sense of accomplishment from completing a 5K race. But he or she can feel a much more significant sense of accomplishment from completing a marathon. 11 years ago I set a lifetime goal for myself to walk 25,000 miles, the equivalent of a walk around the world. I created a Walk Around the World website to help me track my progress toward that goal (the site is also helping almost 300 other people set and reach their walking goals). At my starting rate of 600 miles per year, I set myself a goal that I could not complete for at least 40 years. Will I feel a big sense of accomplishment 30 years from now when I reach that goal? You bet I will! And you will feel a greater sense of achievement in participating in a seven-year-long project than you will participating in a seven-month-long project.
5. Your participation matters. You may only be one of several thousand participants in OGR-27. You may only have one CPU to contribute, compared to another participant’s server farm with hundreds of CPUs. You may only participate in the project for one week each year. But each work unit you complete brings the project a few more minutes or hours closer to completion. Each work unit you complete rules out several billion possible rulers. Each work unit you complete has a chance of finding the optimal ruler, of creating brand new knowledge.
It may be hard to get excited about a project which won’t end for seven years, but that shouldn’t discourage you from participating. My computer is currently testing about 55.6 million OGR-27 rulers every second. While I wrote this entry the computer tested over 400 billion rulers. Those rulers are an almost immeasurably small percentage of the total rulers which need to be tested. They only removed two hours of computing time needed to complete the project. But, except for a double-check by another participant, those rulers won’t have to be tested again. It wasn’t exciting. It is satisfying.