parrt-rules
List of Known parrt Rules
(Many of these rules are common in computer science; parrt just keeps repeating these)
| Parrt Rule # | |
|---|---|
| 1 | Stop what you're doing right now and buy a backup drive or mozy/backblaze/carbonite |
| 2 | Be offended by bad software |
| 3 | Be tenacious |
| 4 | Be courageous |
| 5 | Learn to use the command-line |
| 6 | Use revision control; you can even keep track of which versions of your machine learning model are best, using the commit message to describe the hyper parameters and achieved metrics |
| 7 | Learn to use a debugger and profiler |
| 8 | Practice separation of concerns; e.g., don't merge the loading of documents and word vectorization code, unless speed is critical or it's actually easier to do it that way. Separation of concerns allows you to make reusable functions |
| 9 | Use the principle of least astonishment; do what other programmers will expect |
| 10 | If you can make something work in only one way, the correct way, to do that |
| 11 | Only reality matters; what you think is happening is totally irrelevant. What the data and the code be your guide |
| 12 | When your code is doing something really weird, such as not executing a print statement that it should be, verify you are running the code you think you are |
| 13 | The computer is doing exactly what you tell it to do. There is no mystery |
| 14 | Never upgrade your OS unless you must because you will likely break something, particularly in open-source software that you use. Upgrade "point releases" and security fixes, but not, say, 10.13 to 10.14 unless you absolutely have to. |
| 15 | Don’t solve problems until they exist |
| 16 | Never let the computer win; if you decide to fix a bug, never give up |
| 17 | Reduce number of moving parts to avoid a point of failure |
| 18 | Automation helps productivity, yes, but it's also critically important to making things work correctly and reproducibly |
| 19 | Use the simplest possible thing that will work; only sacrifice readability if performance is important |
| 20 | Precision does not equal to accuracy |
| 21 | Effort is not the same thing as progress |