Six Degrees of Computation | |||||||||||||||||
This idea would require a simple computer program that displayed a screen asking 3 questions - 1. What equation/formula are you interested in? 2. What constant do you want replaced? 3. How many degrees of separation do you want? After you answer the questions the program would search a database that had been pre-loaded with as many standard equations / formulae as possible (engineering books are full of them) looking for ones that said the chosen constant equaled something else. The first equation it found would then be used as a replacement value in the formula you chose. For example, if E=MC2 was the formula you picked and C was the constant you wanted replaced then the C might be replaced by f*w (frequency times wavelength). This would be the first degree of separation. Next, each constant in the replacement value (f and w in our example) would be replaced with equivalent values found for them. That would be the second degree of separation. This process would continue until you got the number of degrees of separation you asked for. Then the result would be displayed for review. Some new and surprising connections would surely come up because the program would in effect be a divergent thinker (someone who can take two ideas from seemingly unrelated disciplines and connect them to form a new idea). The following two rules would have to be enforced in order to make this program work - 1. Each constant must always mean the same thing: m=mass, c=speed of light, etc.. 2. No replacement formula could have a constant that had already been used in a previous formula (preventing recursive logic). Additional Notes: The computer program's choice would be based on on its sequential reading of the formulae in the database. A default setting of one would make it pick the first valid replacement it found. That could be changed though just by adding another question to the prompt screen that asked the user if they wanted to replace the default selection criteria with a different number. If the user keyed a 5 into that field then the program would always pick the 5th replacement formula it found. Or a Y/N "random" flag could be added to the prompt screen that would make the program pick a random valid replacement in the sequence of possibilities when it was changed to a "Y". Or a Y/N "common" flag could be added to the prompt screen that would make the program pick a replacement whose constants were found most ofen in the database when it was changed to a "Y" (increasing the odds that whoever was reading the results would be familiar with the values found in them).
treadair, Aug 10 2004
What do you think of this idea or comment? | |||||||||||||||||
Users who liked this idea also liked: | ||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||
Add your comment
I think this type of "equation mining" is unlikely to produce novel results given that equations don't adequately describe in many cases the underlying nature. Differential equations such as maxwell's equations may be a solution to this, however in many cases to see the similarities between disparate science areas, approximations and partial solutions to these differential equations need to be made. Computers are very bad a pattern matching, people are very good, I think that's why this isn't done.
This is a cool idea but in the end it is likely to produce the ultimate equation which is one = one;