When comparing Docker vs Cygwin, the Slant community recommends Cygwin for most people. In the question“What are the best power user tools for Windows?” Cygwin is ranked 18th while Docker is ranked 26th.
Ranked in these QuestionsQuestion Ranking
Pro Allows for portable application deployment
Docker creates a single object, containing an application with its dependencies, that can be moved between any docker-enabled machines, guaranteeing the same environment for application execution.
Pro Git-like capabilities
Docker tracks changes in systems. It allows for commits and rollbacks and for quick deployment due to having to deploy only the updated code.
Pro Allows re-using components
Docker essentially allows creating boilerplate systems (a LAMP stack, for example) that can be used as a starting point on multiple projects. And you can find multiple such containers already created by people in their public registry.
Pro Automatic build
Allows automatically assembling a container from its source code.
Pro Provides easy sharing and installation of containers through a public registry
Docker allows easily pushing and pulling containers to and from their public index.docker.io registry. Additionally, dotCloud maintains a list of official repositories of the more popular containers.
Pro Works in virtualized environments
You can set up Docker within an already virtualized environment such as a virtual machine. This allows you to run Docker on Mac and Windows, among other use-cases.
Pro Low overhead
Pro Supports a wide range of isolation tools
Docker can be used with OpenVZ, systemd-nspawn, libvirt-lxc, libvirt-sandbox, qemu/kvm, BSD Jails, Solaris Zones, and chroot.
Pro Tool ecosystem
Pro Additional packages available via Cygwin Ports project
Pro Better default terminal
Cygwin's default terminal "Mintty" is far better than CMD.
Pro You can run graphical applications
It has X Server.
Con Large image size
Con Security concerns
Con Kernel OS fragmentation
Con Many incompatibilities
A lot of tools don't actually work with Cygwin, and documentation is often sparse, so for many use cases, it's worth the small bit of extra effort to just run a full virtual machine.