This article is relatively messy, the thinking is not very clear. The content of this book is mostly about how to calculate the performance of a programmer, and compare it to the athlete's record in a sporting event. Personally feel this book if the project manager to read or personnel reading will have more feelings. I'm just about to take a glimpse of the method and extract the knowledge that I think is more useful.
A simple and successful team model
Mode 1 There is a person who does not regret doing trifles.
Mode 21 to two cattle people raise the level of the whole team, although the Bulls do not have a corresponding title
Mode 3 often appears in small business teams, and the project hits the wall at 80%. There is always a person in the team that can survive to inspire morale.
Two
The author of a team, the software from version 1 to part 1.1 took 6 months. And did not add any new features. It took only 9 months from 1.1 to 2, and there were many enhancements. The main change of the team was that the original team left two programmers who were particularly dependent on others, and the new programmers could work on their own and find out the bugs in the software themselves. So the author thinks that a team must have a variety of abilities to be successful.
Three
The outliers and outliers of the evaluation data are not in the normal range, such as the rapid decline of the workload. It can also be an unexplained point, such as someone with a poor academic background, but the efficiency is very high.
Anomalies may be caused by accidental events such as physical discomfort. can be ignored.
Outliers may appear on the surface of the heterogeneous, is the people who break the routine. Ignoring outliers can limit our understanding of success models, such as the Black Swan effect.
Four
Peak and valley values represent cycles.
When you are studying, you should be aware of the project.
1 of statistical projects have limitations, and many useful items are not counted.
2 The value of the statistic is not the same as expected. For example, the author found such a phenomenon. People on the team are often sold, and technical support interrupts the quality of the project, which is contrary to the common belief that programmers should not be disturbed.
Criteria for selection of six statistical projects
1 Easy access to data
2 Easy for non-programmers to understand. This facilitates management, manpower and other communication
Seven how to evaluate a programmer. We can be seen here as how to develop our own abilities.
1 Core Responsibility Performance
2 Code Test Quality
3 How many areas can be covered
4 Take the initiative to solve their own problems. Proactively identify other people's problems
5 Innovation
6 Handling pressure
7 Adversity
8 Interacting with others
9 leadership, support for teammates, ability to guide others
10 understanding of the project and ability to accept team roles
Evaluation criteria for eight teams
1 user response to the new version. such as the adoption rate of the new version
2 How to compare with competitor software
3 Quality
4 new version Delivery rate
Nine the author gave an example when he was speaking to the team.
Two teams, a team formed when the members of the background are very good, and know each other. Second, the background of another team is not very good, people are also temporary formation. But the latter team was successful and the previous team failed. The author has a few gains
1 successful team complex tasks are concentrated in a small number of hands. Other programmers take on a lot of work, but not hard. The team that failed is almost everyone's workload is similar, the average complexity is high.
2 successful teams each work in multiple areas. Failed teams each work in a very small area
3 Successful team of people innovation, also initiative.
"Programmer's metrics, improving software team Analytics" Reading notes
