Understanding Package Managers: pnpm vs npm vs yarn

You should probably know by now that package managers are crucial tools in the JavaScript ecosystem, automating the process of installing, updating, configuring, and removing project dependencies. I will try to provide an in-depth comparison of three popular package managers: pnpm, npm, and yarn, explaining their inner workings, key features, and practical implications for developers.

First, let should know that this package manager have the same functionality, but they have different ways on approaching them. And we will be looking at them.

npm (Node Package Manager)

First lets talk about NPM (Node Package Manager), this package manager is the default package manager for Node.js, which is a runtime environment that enables the execution of JavaScript code on the server side, outside of a browser. You all probably know npm because almost all beginners and learners learned about npm when you started. Additionally, NPM enables automation of tasks like running tests, building projects, or deploying code through custom scripts defined in the package.json file. It's an essential tool in the JavaScript ecosystem, particularly for Node.js development, making it easier to manage and share reusable code.

How npm works:

1. Dependency Resolution:

   • npm reads the package.json file to determine project dependencies.
   • It constructs a dependency graph, resolving version conflicts using a deterministic algorithm.

2. Installation:

   • npm installs packages in a nested structure within the node_modules folder.
   • Example structure:

node_modules/ 
├── package-a/ 
│ └── node_modules/ 
│ └── package-b/ 
└── package-c/

1. Flat Structure:

   • npm v3+ attempts to flatten the dependency tree to reduce duplication.
   • This can lead to "dependency hell" where different versions of the same package are required.

2. Package Lock:

   • Uses package-lock.json to ensure consistent installs across environments.
   • Contains the exact version of each package in the dependency tree.

3. Scripts:

   • Allows defining custom scripts in package.json.
   • Example:

"scripts": { "start": "node server.js", "test": "jest" }

Pros:

• Largest package ecosystem with over 1.5 million packages
• Built-in with Node.js
• Extensive documentation and community support

Cons:

• Slower installation compared to yarn and pnpm
• Can lead to large node_modules folders (sometimes jokingly referred to as "node_modules black hole")
• Potential for dependency conflicts

yarn

Yarn is a package manager for JavaScript that was developed by Facebook in collaboration with other companies, as an alternative to NPM. It aims to improve the speed, reliability, and security of dependency management in JavaScript projects. Yarn enhances performance by using a cache to store downloaded packages locally, which speeds up subsequent installations. It also ensures consistency across environments by generating a yarn.lock file that locks down the exact versions of dependencies used in a project, preventing discrepancies between different setups. Additionally, Yarn offers better offline support, more predictable and deterministic installs, and improved security by verifying the integrity of downloaded packages. These features make Yarn a popular choice for managing project dependencies, particularly in larger or more complex JavaScript projects.

How yarn works:

1. Dependency Resolution:

   • Like npm, yarn uses package.json for dependency information.
   • Implements a more sophisticated resolution algorithm to handle complex dependency graphs.

2. Parallel Installation:

   • Installs packages in parallel, significantly improving speed.
   • Uses a global cache to store downloaded packages, reducing network usage.

3. Offline Mode:

   • Caches packages for offline use.
   • Can install dependencies without an internet connection if they're in the cache.

4. Deterministic Installs:

   • Uses yarn.lock for consistent installations across different machines.
   • Ensures that the same dependencies are installed regardless of install order.

5. Workspaces:

   • Supports monorepo structures with workspaces.
   • Example package.json for a workspace:

{ "private": true, "workspaces": ["packages/*"] }

Pros:

• Faster than npm, especially for large projects
• Reliable and consistent installations
• Enhanced security features (checksum verification)

Cons:

• Still creates large node_modules folders
• Some features require using Yarn-specific commands

pnpm

pnpm is a fast, disk space-efficient package manager for JavaScript that is an alternative to NPM and Yarn. It is designed to improve performance and save disk space by creating a single store of packages on your computer, instead of duplicating dependencies across multiple projects. When you install packages with pnpm, it creates hard links to the shared store, making the installation process faster and reducing the overall disk space used.

pnpm also ensures that dependencies are strictly isolated, which can prevent potential conflicts and issues in your projects. This strictness helps maintain consistency and reliability, particularly in complex projects with many dependencies. Additionally, pnpm supports features like workspaces, allowing you to manage multiple related projects within a single repository. Its efficiency and focus on performance make pnpm an attractive choice for developers looking to optimize their development workflow.

How pnpm works:

1. Content-Addressable Storage:

   • Stores all packages in a global store, typically located in ~/.pnpm-store.
   • Each project links to this store instead of having its own copy of packages.

2. Symlinks:

   • Uses symlinks to create a nested node_modules structure.
   • Example structure:

node_modules/ 
├── .pnpm/ │ 
├── package-a@1.0.0/ 
│ └── package-b@2.0.0/ 
├── package-a -> .pnpm/package-a@1.0.0/node_modules/package-a 
└── package-b -> .pnpm/package-b@2.0.0/node_modules/package-b

1. Efficient Storage:

   • Only one copy of a module version is saved on disk, regardless of how many projects use it.
   • This can save gigabytes of disk space for large projects or multiple projects on the same machine.

2. Strict Mode:

   • Prevents packages from accessing arbitrary packages in the node_modules folder.
   • Ensures that only declared dependencies are accessible, improving security and preventing "phantom dependencies".

3. Monorepo Support:

   • Native support for monorepos without additional tools.
   • Example pnpm-workspace.yaml:

packages: - 'packages/*'

Pros:

• Dramatically saves disk space
• Fast installation and updates
• Ensures package isolation and prevents phantom dependencies
• Built-in monorepo support

Cons:

• Less widely adopted compared to npm and yarn
• May have compatibility issues with some tools expecting a traditional node_modules structure
• Learning curve for developers used to npm or yarn

Comparison Summary

1. Installation Speed:

   • pnpm > yarn > npm
   • pnpm and yarn are significantly faster than npm, especially for larger projects.

2. Disk Space Usage:

   • pnpm > yarn ≈ npm
   • pnpm can save up to 80% disk space compared to npm for projects with many dependencies.

3. Ecosystem & Adoption:

   • npm > yarn > pnpm
   • npm has the largest ecosystem, but yarn and pnpm are gaining popularity.

4. Dependency Resolution:

   • All three use similar algorithms, but pnpm's approach is unique and more efficient.

5. Lock File:

   • All use lock files for consistency (package-lock.json, yarn.lock, pnpm-lock.yaml)
   • Lock files ensure reproducible builds across different environments.

6. Monorepo Support:

   • pnpm > yarn > npm
   • pnpm and yarn have built-in support for monorepos, while npm requires additional tools.

7. Security:

   • pnpm > yarn > npm
   • pnpm's strict mode and yarn's checksum verification provide additional security layers.

Practical Implications

1. Project Onboarding:

   • npm is often the easiest for new developers due to its ubiquity.
   • pnpm and yarn may require additional setup but can significantly improve project efficiency.

2. CI/CD Performance:

   • pnpm and yarn can dramatically reduce build times in CI/CD pipelines due to their faster installation and caching mechanisms.

3. Disk Space in Docker:

   • Using pnpm can significantly reduce Docker image sizes for Node.js applications.

4. Large-Scale Development:

   • For large projects or organizations working on multiple projects, pnpm's space-saving feature can be a game-changer.

5. Monorepo Management:

   • pnpm and yarn are better suited for managing monorepos without additional tools.

My Take

While a lot of you uses npm and yarn. Me and a lot of developers is moving to pnpm. The main reason is not only that its fast, but it also does not eat a lot of your storage. For me, that is the very main thing why I started using pnpm. If you think different than I am, please comment down bellow. Let me know what you guys think.