Transfer: https: //mp.weixin.qq.com/s/acuQZPPFGVzEiWQYpe5VsA
1. reconstruction is the main programmer skills.
2. The logbook can improve brain capacity.
3. The first use profiler survey, only face to talk about optimization.
4. Comment expensive, not much. Eliminate aunt-like "cases of note." Mountains and plains of Sui Suinian notes, is in fact the background noise.
The programmer of ordinary + google = super programmer.
6. The test unit is always advantageous.
7. Do not write to write framework to achieve. Preferably, in turn, extracted from the framework of the prototype.
8. clear code structure, other issues are not a thing.
9. hardcore style of good projects, a key test, one-click publishing, a key deployment; rotten wretched nature of the project, word of mouth, can not stand text, secretive.
10. Do not be afraid coding change, to embrace change.
11. Chang charging. Programmers only one die: earth dead.
12. programming things, isolation is the direction, the key is named, is the protagonist testing, debugging supplement, version control is regret.
13. The line of code a soldier. Organizational form in order to have combat. The size of the unit should not be too large, thousands of classes, people row easily into mass graves.
14. The reconstruction / optimization / repair Bug , only for a while.
Simple encapsulation module 15. Note that, attention layered complex modules.
16. The human brain is limited performance, clean and better than clutter. I can not read the code attempts to organize the format; bad interface used, the attempt to reseal.
17. iterative speed determines the intensity of work. Want to do more with less, we begin to simplify the development process, accelerate the speed of iteration.
18. Forget about writing code optimization, premature optimization is equivalent vandalism; forget the code for optimization, to optimize based on performance testing, rather than entangled in the lines.
19. The best tool is a pen and paper; secondly is good markdown.
20. leader asked the task time, if not answer, task split may be enough fine.
21. would rather count a week essential estimate day. Too "optimistic" easy boss frightened.
22. The most useful language is English. Secondly it might be Python.
23. Seeing is believing. Draw results at a glance. Will greatly shorten the time-consuming debugging.
24. resources, the code should work under version management. Resource matching error code matches the error is more difficult than investigation.
25. Do not imagine based development, to be based prototyping. The prototype is quick verification value ideas to help you save time.
26. The sequence of the preferred clear text. Such as binary, confusion, encryption, compression, and so on together when necessary.
27. The compiler is always better than you know micro-optimization. It does not only strive toward good.
28. Do not set too high, too far, too small plans. Even given no avail.
29. At least half of the time will be spent on integration. Time, time, time is always not enough.
30. The mainstream opinions / methods / style / habits contrary, after a review of their most reliable.
31. The emergence of Bug initiative investigation, whether it is not yours. This service allows you the ability to soar, soaring personal image; if your Bug pulling out others ..... Oh, then you will be very passive ~ ≧ ﹏ ≦
32. I do not know how to pick the thin election technology books. At least not too expensive, and you can read.
33. The git is the best. Simple, reliable, and free.
34. The only "predictably irrational" throw assertion.
35. The the Log write time and classification. And to be able to redirect the output.
36. Notes is somewhat less documentation. Better yet clear name. Let the code tell their own stories.
37. create the wheel is good exercise method. Provided that you have seen the other wheels.
38. A code Review preferably in group / node pairs. Have a certain understanding of the business, proposals will be more valuable (but not absolute). And will not become a burden. Administrators personal review it is easy to become a bottleneck team.
39. The question before the research to do first. He asked not to be despised on both points, and a waste of their time.
40. Do not ever underestimate the procedure Yuan (╯3╰)
