A few years ago, I ran a cloud rendering startup that worked with VFX studios.
We worked with a few mid-size studios, that used Windows on all their machines. We were faced with the challenge of deploying Windows server farms, and find practical ways to install programmatically heavy software (3dsMax, Maya, V-ray…) on all the servers. Don’t forget that at the time Windows didn’t have SSH, and scripting was mainly Powershell.
We were looking for something similar to Docker, but for Windows software. We finally came across Spoon, a company that proposed a software to package any Windows application into a self-contained executable.
The software basically started specific drivers that performed file system and network redirection, so that the binaries executed by it would “see” their dependencies like if they were at their nominal location while in reality they would be packaged into the super-executable created by Spoon. You just had to activate Spoon, run your software installer inside the Spoon sandbox, and tell Spoon where to package all the installed stuff.
You could then copy this executable from machine to machine and run your software anywhere without worrying about dependencies. What is amazing is that this discrete company had designed in the 2000’s the “equivalent” (conceptually) of Docker, but for Windows! And from what we could see, it worked really well. But their achievement triggered no hype, and their tech remained quite confidential.
A couple of years ago, they even open sourced their tech under the name Spoonium, probably influenced by the Docker success story. It seems it didn’t take off, and after yet another rebranding, the project is now closed source again, and the company is now named Turbo.
It never ceases to amaze me how impressive and truly useful tech can sometimes miss fame, just because of platform (Windows) and timing (containerization was not a trending topic in the 2000’s).