Intro to Unix and shell

Reference

Great (paid) LinkedIn course by Kevin Skoglund “Unix essential training”

What is Unix?

• an OS developed by AT&T employes at Bell Labs in the 1960s/70s
• in 1972 it is rewritten in C programming, which made it portable. It then spread outside AT&T
• it still powers devices all around
• you can use it from the Command line or from GUI interfaces

What are the kernell and shell?

• kernel is the core of the OS. it allocates time and memory to processes
• shells is the outer layer of the OS. it sends requests to the kernel
  • you can choose from different shells (sh, bash, zsh, etc.)

Fundamental

• Ctrl + C = Cancel/stop the current command (SIGINT)
• Ctrl + Z = Pause current command (suspend, SIGTSTP)
  • fg Resume the last paused command
• Ctrl + a = start of the command
• Ctrl + e = end of the command
• q = close dialogue
• Ctrl + L → clears the terminal screen (but not history)

Mac is a “Unix system”

The macOS is a Unix-native system, i.e. it is built on a Unix core (Darwin). This enables you to use Unix commands in the terminal, as opposed to Windows, which is built on a different core.

Why Mac is preferred for programming?

• Mac has a better terminal emulator (Terminal, iTerm2) than Windows (Command Prompt, PowerShell)
• MacOS is more stable and less prone to viruses than Windows
• Mac with chip Apple Silicon (M1, M2, M3) has better performance and energy efficiency than Windows laptops with Intel or AMD processors
• Mac is a Unix system, so you can use Unix commands in the terminal (e.g. ls, cd, mv, cp, rm, etc.)
  • [On Windows, to run Unix commands, you need to install a Unix-like environment, e.g. 
    • WSL (Windows Subsystem for Linux)
    • Cygwin (a large collection of GNU and Open Source tools which provide functionality similar to a Linux distribution on Windows)
    • Git Bash](an application for Microsoft Windows environments which provides a bash emulation used to run Git from the command line.)

MacOS and open source software

• MacOS is more compatible with open-source software and tools used in data science (e.g. R, Python, etc.)
  • On Windows, you may face compatibility issues with some open-source software and tools.
• Many libraries and packages used in data science are developed and tested first on Unix/Linux systems, so they may work better on MacOS than on Windows.
• Some dependencies and tools may not be available or may require additional configuration on Windows (Rcpp, devtools, spacy, etc.)
• Homebrew, which is a popular package manager for MacOS (& Linux), makes it easy to install and manage open-source software and tools (es. git, python, r, node, wget, ecc.). Plus, it avoids having to download different installers, configure PATHS, manage version conflicts etc.
  • Windows alternatives would be Chocolatey or winget, but they are not as widely used or supported as Homebrew.
• Docker, which is a popular containerization platform used in data science, also works better on MacOS than on Windows, as it uses a Linux kernel to run containers.

Table 1: Comparison of macOS/Unix vs Windows for Data Science Workflows

Area	macOS_Unix	Windows_Standard	Windows_Positron
Shell / PATH	Single Unix shell (zsh/bash). PATH always coherent.	Multiple shells (CMD, PowerShell, Git Bash, WSL). Frequent PATH conflicts.	Positron manages interpreters better, but PATH issues remain.
R & Packages	CRAN install + compilation works out of the box.	Requires Rtools for many packages → extra step, errors likely.	Still requires Rtools. Positron smooths management but OS issue remains.
Python & Envs	Native support. venv/conda integrate with terminal.	Multiple installs common. PATH confusion across shells.	Positron can choose envs in IDE, but conflicts outside remain.
Git / GitHub	Git native via Xcode CLI or Homebrew. One SSH dir (~/.ssh).	Git for Windows + Git Bash + WSL = multiple key locations, SSH mismatches.	Positron integrates Git/GitHub login, centralizing SSH inside IDE.
Docker / Containers	Native support. ARM64 stable on Apple Silicon.	Needs WSL2/Hyper-V → overhead, resource heavy.	Still needs WSL2/Hyper-V. Positron smooths workflow but not overhead.
IDE	RStudio, VS Code, PyCharm run natively with Unix shell.	RStudio and VS Code work, but rely on Git Bash/WSL.	Positron unifies R, Python, Quarto, Git → single IDE lowers friction.
Maintenance	Minimal IT support needed. Easy updates via Homebrew.	Frequent issues with PATH, Rtools, Docker/WSL, SSH keys.	Less IDE friction, but IT still needed for OS-level issues.

Table 2: Comparison of macOS/Unix vs Windows for Data Science Workflows

Area	macOS_Unix	Windows_Standard	Windows_Positron
Shell / PATH	Unica shell Unix (zsh/bash). PATH sempre coerente.	Più shell (CMD, PowerShell, Git Bash, WSL). Conflitti PATH frequenti.	Positron gestisce meglio gli interpreti, ma i problemi di PATH restano.
R & Pacchetti	Installazione da CRAN + compilazione funziona subito.	Richiede Rtools per molti pacchetti → passaggio extra, errori possibili.	Rtools serve comunque. Positron semplifica la gestione, ma il problema rimane.
Python & Ambienti	Supporto nativo. venv/conda integrati col terminale.	Spesso più installazioni (system, Anaconda, Store). Conflitti PATH tra shell.	Positron può scegliere l’ambiente in IDE, ma i conflitti fuori restano.
Git / GitHub	Git nativo via Xcode CLI o Homebrew. Una sola directory SSH (~/.ssh).	Git for Windows + Git Bash + WSL = directory multiple, conflitti SSH.	Positron integra Git/GitHub con login centralizzato alle chiavi SSH.
Docker / Container	Supporto nativo. ARM64 stabile su Apple Silicon.	Richiede WSL2/Hyper-V → overhead, consumo risorse elevato.	WSL2/Hyper-V servono comunque. Positron semplifica i workflow ma non l’overhead.
IDE	RStudio, VS Code, PyCharm girano nativamente con shell Unix.	RStudio e VS Code funzionano, ma dipendono da Git Bash/WSL.	Positron unifica R, Python, Quarto, Git → un solo IDE riduce gli attriti.
Manutenzione	Minimo supporto IT. Aggiornamenti facili con Homebrew.	Problemi frequenti con PATH, Rtools, Docker/WSL, chiavi SSH.	Meno problemi lato IDE, ma l’IT serve ancora per i conflitti del sistema operativo.