Xmonad vs dwm

XMonad is a very minimal and efficient window manager, especially if the user is familiar with Haskell.

XMonad is written, configured, and fully extensible in Haskell. This means that users aren't limited to a small set of pre-programmed layouts and actions: anything can be programmed into the configuration.

It's simple to modify basic settings, and the example config has lots of comments to get you started. Haskell keeps this code clean, concise, and readable, and its type system keeps you safe from any serious mistakes. This makes it fast and light, even on very small and slow systems.

Use of Haskell, in conjunction with smart programming practices, guarantees a crash-free experience.

The entire window manager is extremely small, and includes nothing beyond basic window manipulation and tiling. Out of the box, there are no window decorations, status bar nor icon dock; just clean lines and efficiency.

If you enjoy programming, you can even add features to XMonad to make it your perfect desktop environment, and the Contrib modules give you most of what you need to do exactly that.

The use of Haskell as an extension language means that popular pieces of functionality are easily shared and widely available as Haskell Libraries. Many default layouts, and tools for quickly and easily building your own, are available through XMonad-contrib, and highly re-usable configurations are commonly shared through blog articles and the Xmonad Wiki. The documentation in XMonad-contrib is very clear and easy to read.

XMonad has full support for Xinerama: windows can be tiled and managed across multiple physical screens.

Configuration is compiled into the WM, and it can be changed/updated on-the-fly, without requiring a full reload.

XMonad can handle multi-monitor setups by default.

XMonad uses dynamic tiling which means that it automatically handles arranging your windows into various layouts which the user can cycle through.

XMonad separates screens and workspaces. A screen "projects" a workspace. You can put a window to a specific screen, regardless of which workspace is currently projected onto that screen. This is more intuitive than other WMs e.g. i3, which only has the notion of workspace but not "screen" and requires you to remember workspace numbering. It is especially beneficial for multi-monitor setups.

Dwm is part of the suckless suite of tools, and encourages users to extend and configure it by modifying the code itself. To this end, dwm is kept under 2000 SLOC, and is an exemplar of clean, readable code (C). This, while giving users all the flexibility they could ask for, also makes dwm as lightweight as possible, and means that users have a full understanding of how it works.

Dwm is a low-resource window manager that is entirely simplistic in design.

An example of this is the application of alt-tab to switch between two tags.

Dwm's design paradigm is to use tags to group clients (applications) that can then be pulled into a view (workspace); this allows you to view multiple clients at once and to assign or reassign those tags and their related views on the fly.

Contrary to most other window managers, when you view a tag you are not ‘visiting’ a workspace: you are pulling the tagged windows into a single workspace.

Combined with rules in the config.h, this makes for a flexible and responsive means to manage your workflow.

Configuring dwm is straight-forward thanks to its config.h file (though it will have to be rebuilt for the effects to take place).

Dwm has support for XRandR and Xinerama, allowing for multi-monitor support.

The dwm status bar can be set to display all kinds of useful information, such as volume level, wifi signal strength, and battery notification.

Thanks to the small codebase, many users contributed patches to the suckless website. They offer unique functionality, e.g. swallow or fakefullscreen, that is not seen in many other WMs.

Uses Master&Stack layout by default.