1. The computer is not as accurate as imagined

Fundamentally speaking, a computer is a machine that uses a potential difference to perform calculation operations. The high potential is 1 and the low potential is 0. Only 0 and 1 are required for calculation. All digital machines follow this principle.
Can the binary system represent the decimal system 0.1, 0.1, 0.3, 0.4, 0.6? Can't!

2. Avoid floating-point arithmetic when accurate calculations are required

The characteristics of floating-point numbers make it impossible to perform accurate calculations, that is to say, it always has errors.
The float in C language can only guarantee the accuracy of 6 decimal places after the decimal point
The double can only guarantee 15 digits
If you calculate 1/3+1/3, the answer given by the computer is 0.666667 (float type)
But it's actually 2/3
Therefore, when we need accurate calculations, we should use integers instead of floating-point numbers.

The speed of double operation and float operation is almost the same
Since the speed difference is not obvious, considering the accuracy, we should use double type arithmetic

In the PC operating system, the size of the int type in a 32-bit system is 4 bytes
The integer size is 2 bytes in a 16-bit system
The size of int type in 64-bit operating system is 8 bytes
Of course, their respective ranges are also different
Therefore, we should develop a good habit of confirming the size of the data type before coding, and it is best to mark the size of the data type next to the variable

int number = 7000; //int型大小为4字节

If the variables in the program have practical meaning, you have to give their units, otherwise who will understand

sum = sum +(num / 1000); //num以元为单位，sum以千元为单位
total = total + (sum / 1000); //sum以千元为单位，total以百万元为单位

When long bytes are converted to short bytes, precision is lost
The way to prevent implicit conversion is to add a prefix that can represent the data type before the variable name, as follows:

double dpi = 3.14159;
long int liNum;
long int liRad;
......
liNUm = dpi * liRad * liRad;

After declaring and using these variables, you can discover the possible multiplication or assignment operations between variables of different data types at the first time

The butterfly effect also applies to computer operations
The computer is not what we think
After repeated calculations of floating-point numbers, they may have deviated from our expectations. These numerical deviations have led to shipwrecks, crashes, and explosions in power stations. . .
Therefore, if you need to repeatedly calculate subtle numerical differences in the process of coding, designing program algorithms, and designing system frameworks, please check the results of the calculations many times! ! ! !

There are many details in this article, mostly from the variable aspect. Take a look around, it's actually pretty good.