Note:
Recently I thought about how to make data, although there is no operation, but would like to share with you the method.
method 1:
Hard to think. . . Suitable for small data, according to the code error-prone may want some of their own data, this is not to say, like after understanding the code.
Method 2:
Random data, this will be no size limit, and big data can do, but the problem again, I do out of the water data is likely not entirely random, then some random topic even only a kind of output, and not good screening. So, this is the solution I think: to beat, Helena has the details of the error code and the correct code for the shoot, and then screened to take the complexity of the code in question and normal codes to shoot, to find out (within t Code record time, over the time limit is not output, to take a normal write on the line) data for screening.
Method 3:
The purpose of random, random add some conditions to run the process, the less random random, random direction we want, matched filtering method 2, and method to do a few small data just fine.