diff --git a/README.md b/README.md
index 68f6230..4aa3ae0 100644
--- a/README.md
+++ b/README.md
@@ -1,2 +1,13 @@
-# system-programming
-A repo for my experiments on system programming
+# 📚 Linux System Programming – Chapter Index
+
+Welcome! This is a collection of chapter summaries from the book **Linux System Programming** by Robert Love. Each chapter is organized in its own folder.
+
+## Available Chapters
+
+- [Chapter 1 – Introduction and Core Concepts](chp1/README.md)
+- [Chapter 2 – File I/O](chp2/README.md)
+- [Chapter 3 – Buffered I/O](chp3/README.md)
+- [Chapter 4 – Advanced File I/O](chp4/README.md)
+
+> Each file contains an English summary of the chapter's key concepts.
+
diff --git a/chp1/README.md b/chp1/README.md
new file mode 100755
index 0000000..fcc56e6
--- /dev/null
+++ b/chp1/README.md
@@ -0,0 +1,28 @@
+# Chapter 1 – Introduction and Core Concepts
+
+## 🔧 What is System Programming?
+- Low-level programming that interacts directly with the **kernel** and system libraries via **system calls**, typically written in **C**.
+- Includes shell tools (bash, gcc, coreutils), daemons, webservers, etc.
+
+## Why Learn It?
+- Even in high-level environments (JS, Java, Python), understanding internal mechanisms leads to better, more efficient code.
+- Still essential for core components of Linux systems: kernel, drivers, glibc, gcc.
+
+## Core Elements
+1. **System calls** – entry points from user process to the kernel (e.g., `read()`, `write()`).
+2. **C Library** (glibc) – interface between user programs and system calls.
+3. **C Compiler** (gcc) – handles APIs, ABIs, and object formats.
+
+## API vs ABI
+- **API**: interface at the source code level (e.g., libc functions).
+- **ABI**: binary layout and conventions (e.g., syscall invocation, parameter ordering). Defines compatibility between systems and architectures.
+
+## Standards and Compatibility
+- POSIX/SUS as a reference, but Linux is not officially certified.
+- The **Linux Standard Base (LSB)** initiative extends POSIX/SUS to ensure binary compatibility across distributions.
+
+## Linux Basics
+- **Files**: name + inode, represented by **file descriptors** (int). Linear byte streams (no formatting).
+- **Processes**: entities with address space, pid, users, groups, permissions, signals.
+- **IPC**: inter-process communication mechanisms (pipe, FIFO, socket, shared memory, semaphores).
+- **Error handling**: functions return values and set `errno`. Useful functions: `perror()`, `strerror()`, `strerror_r()`.
diff --git a/chp2/README.md b/chp2/README.md
new file mode 100755
index 0000000..94affd1
--- /dev/null
+++ b/chp2/README.md
@@ -0,0 +1,30 @@
+# Chapter 2 – File I/O
+
+## Basic Concepts
+- File I/O is central in Linux: every entity (file, socket, device) is treated as a file.
+- The kernel manages a **file table** per process, with descriptors and metadata like offset and mode.
+
+## Core Operations
+1. **`open()`**:
+   - Opens a file with flags (O_RDONLY, O_WRONLY, O_RDWR, O_CREAT, O_SYNC, O_DIRECT).
+   - Returns a file descriptor or -1 on error.
+   - Sets owner and permissions for newly created files.
+2. **`read()` / `write()`**:
+   - `read(fd, buf, n)` reads up to **n** bytes, returns bytes read or -1.
+   - `write(fd, buf, n)` writes, handles partial writes, append mode and non-blocking I/O.
+   - Common errors: EAGAIN, EINTR, ECC.
+3. **I/O Synchronization**:
+   - `fsync()` and `fdatasync()` flush to disk.
+   - `O_SYNC`, `O_DSYNC`, `O_RSYNC` for open-time synchronization.
+   - `sync()` syncs all files to disk.
+4. **`lseek()`**:
+   - Moves the **offset** within the file—can be positive/negative; allows sparse writes.
+5. **`close()`**:
+   - Frees the descriptor. Returns 0 or -1 on error.
+
+## Positional I/O
+- `pread()` / `pwrite()` allow read/write at specific offsets without altering the shared file offset (useful in multithreaded environments).
+
+## Multiplexed I/O
+- Functions like `select()` and `poll()` manage multiple file descriptors for asynchronous I/O.
+- More scalable option: `epoll()` (covered in Chapter 4).
diff --git a/chp3/README.md b/chp3/README.md
new file mode 100644
index 0000000..eacadc8
--- /dev/null
+++ b/chp3/README.md
@@ -0,0 +1,26 @@
+# Chapter 3 – Buffered I/O
+
+## Difference Between Byte-Level and Buffered I/O
+- **System calls** like `read()` and `write()` operate at a low level: each call may result in direct disk access.
+- **C libraries** use buffered I/O (`fread()`, `fwrite()`, `fprintf()`), accumulating data in memory to reduce system calls.
+
+## Buffered I/O API
+- Uses `FILE *` type from `<stdio.h>`.
+- Functions:
+  - `fopen()`, `fclose()`: open/close file.
+  - `fread()`, `fwrite()`: binary I/O.
+  - `fprintf()`, `fscanf()`, `fgets()`, `fputs()`: formatted or text I/O.
+
+## Types of Buffering
+- **Fully buffered**: flush when buffer is full.
+- **Line buffered**: flush at end of line (e.g., terminals).
+- **Unbuffered**: no accumulation (e.g., stderr).
+
+## Buffer Management
+- `setvbuf()` or `setbuf()` allows custom buffering.
+- Useful for optimizing performance or interactive environments.
+
+## When to Use Buffered I/O
+- When performance matters and immediate file system synchronization isn't required.
+- For formatted I/O, the C library offers more simplicity and readability.
+
diff --git a/chp4/README.md b/chp4/README.md
new file mode 100755
index 0000000..802516c
--- /dev/null
+++ b/chp4/README.md
@@ -0,0 +1,42 @@
+# Chapter 3 – Advanced File I/O
+
+## File Metadata and the `stat` Family
+- Use `stat()`, `fstat()`, `lstat()` to retrieve file metadata:
+  - File type, size, inode number, permissions, ownership, timestamps (atime, mtime, ctime).
+- The `stat` structure includes all these details.
+
+## File Types
+- Regular file, directory, symbolic link, character/block device, FIFO, socket.
+- Use macros like `S_ISREG()`, `S_ISDIR()` to test file types.
+
+## File Permissions and Modes
+- Represented by 12 bits (e.g., `rwxr-xr--`), plus sticky/SUID/SGID bits.
+- Manipulate with `chmod()`, `fchmod()`, `umask()`.
+- Ownership changed with `chown()` and `fchown()`.
+
+## Links
+- **Hard links**: multiple filenames point to the same inode.
+- **Symbolic links**: separate files that reference other paths; can be broken.
+- Use `link()`, `unlink()`, `symlink()`, `readlink()`.
+
+## Directories
+- Open with `opendir()`, read entries with `readdir()`.
+- Low-level: `getdents()` system call.
+
+## File Timestamps
+- `utime()`, `utimes()`, `futimens()`, and `utimensat()` to update access/modification times.
+
+## File Descriptors and Flags
+- Manipulate with `fcntl()`:
+  - Set non-blocking mode (`O_NONBLOCK`), FD_CLOEXEC flag, file locks.
+  - Duplicate with `dup()` and `dup2()`.
+
+## File Locking
+- `flock()` for advisory locking (whole file).
+- `fcntl()` for POSIX record locking (ranges).
+
+## Temporary Files
+- `tmpfile()`, `mkstemp()` are safe methods to create temp files.
+
+## Summary
+- Chapter 3 focuses on extended file handling: metadata, links, directories, file modes, and advanced descriptor manipulation.