Be it files in a code repository, or an API resource, or a physical machine -- all of them have manuals of usage of some
sort:
- if it's a machine, that operation manual comes in form of a book, help file or reference manual,
- if it is an API, the instructions may come in its schema response,
-
if it's a code repository, it often comes as readme with some Makefile, that implements a kind of callable interface to
the repository itself.
It would be great if we could start using these interfaces in a sort of plug-and-play fashion, that we can use things like
mouse or keyboard.
Here's where "UseMe" standard would come in -- we would choose a general interface description language, and use it
to describe systems and subsystems of all sorts, using the machine-readable UseMe format, that would work like an
abstraction layer on top of those various operation manuals, Makefiles, API schemas, and other type of interface
descriptions.
Just like we have a "readme.txt" for humans, we might want to have "useme.txt", and in fact, the UseMe could even be
integrated as part of readme files, by having a special section with a special tag like %USEME%, that would allow the
developers and engineers inject updated interface specifications into ReadMe itself.
All kind of APIs (GraphQL, REST, SOAP, XML-RPC, etc.) would return not their schema, but the UseMe response.
This would allow people and machines to interface with systems directly with flat learning curve.