Mu Language

The Mu language is an emergent language built for the era of services.


Mu is a new language we’re hypothesising about which is built on the foundations of services rather than any sort of formally defined spec. It is firstly not a replacement for any existing programming language such as Go, java, ruby, python, etc and does not attempt to be comparable to current programming models.

The goal is to lay the foundation for an emergent language where the keywords are services and all we attempt to do is define a grammar to associate these services in a cohesive way that then lets us perform actions at a higher level.


The assumption is that we can steal from the likes of bash while creating something purely human readable or with a fixed type set for interop between services. We should also include the ability to save new programs as the aggregate of many others but in lazy loaded fashion as with haskell execution.

# a service call to the store
# [service] [method] [input] [value]
> store write foo bar

# retrieve the output
> store read foo

# storing the output of such a command
> save result = store read foo

# evaluating the output
> exec result

# passing the output to an input
> input result into broker publish values


The longer term potential for such a language is that it becomes a higher level form of orchestration for services. At a given point when enough services have been written, they become motor functions capable of being combined to perform other forms of execution. Continuing to use existing programming languages to orchestrate these services becomes cumbersome and a dynamic language does not exist to perform operations on services.

By escaping the current programming boundaries we can elevate to a new intermediary representation that can either be leveraged by human beings more easily, spoken as commands or used for other purposes.