I carefully considered various perspectives and finally gave me a reply:
ToValidCodeQuantityMeasure the capabilities of an engineer! Of course, we will not discuss EQ here.
(I suddenly found myself pulling too far)
Advanced or elementary. The final response to his abilities must be through the code they have typed. All their professional abilities must be reflected through the Code:
- How many lines of code are written?
- How long does it take?
- How many changes have been made?
- How many repeated communication and adjustments have been made?
- The total project cycle?
- ...
We set some reasonable standard parameters and then compare them to produce a result:
Amount of valid Code. You don't have to worry too much about the advantages and disadvantages of the Code,ProgramFrom the perspective of normal distribution, all problems will be reflected in the Code through time.
In fact, my point of view may not be correct. This is just the answer I found when thinking about "how to quantify". You may have better answers. I know that most companies have salary or grade standards, what kind of p1-p12, but it does not necessarily meet our needs.
Assuming that the reference system we set up is valid, there is another problem:
How to measure the project's time cycle?
If you are working on a similar project, I believe that your experience can provide a precise reference. What if it is a strange project? The solutions for unfamiliar projects that are estimated by experience have been rejected in the previous section. I have been thinking about this question for a long time. I have also tried to find answers on the Internet and asked a variety of friends. There are someAlgorithm(From IBM). It's complicated. It doesn't seem to be what I need. My friends told me that most of them evaluated through experience. ......
Finally, I affirmed the experience evaluation when I denied the experience evaluation. Of course, it is not the evaluation you think! My thoughts and math
IterationBasically consistent:
- The accuracy of my prediction is evaluated based on the actual completion time from the first project;
- Revise and then evaluate, and then revise
- After multiple corrections, you will get a parameter that can be used to evaluate (judge) The accuracy of my project evaluation.
Assume that it takes three months for me to predict this project, and the number of assessment revisions in the past shows that the coefficient is 1.3 for my assessment. So the project cycle close to scientific accurate prediction is 3*1.3 = 4 months.
I have come up with this solution! Because there is no better or more accurate solution, you are welcome to propose a better solution.. The so-called classification granularity is only one part of my evaluation;
Divide the project into code, add our risk coefficient, and compare the valid code quantity. Finally, a result is obtained.
This result may not be correct or accurate, but in mathematics, we are seeking for extreme values!
This result is obtained after the error of each link is reduced.
It is a trusted reference value !!!
To prove that my opinion is correct, let's look at a mathematical iteration (from # astronomical algorithm #):
-----------------------------------
Calculate SQRT (159) with 8 Valid digits
It is actually the square root of 159. What if there is no calculator?
We know 12*12 = 144, so we can use 12 as the valuation of SQRT (159. Divide 159 by 12 and get 13.2500. Then the arithmetic mean of 12 and 13.2500 is 12.6250, which is a better estimate.
We now use 159 except 12.6250 to get 12.59406, and the average value of 12.59406 and 12.6250 is 12.60953. This value will be more accurate.
In this way, we get:
12 = start value
12.62500000
12.60952971
12.60952022
12.60952022
We can see that 12.60952022 to 12.60952022 are no longer changing, so this is the square root of 159.
Hypothesis> negation> hypothesis> Negation
We are constantly arguing and finally get a result. This result may be what we need!
