Initial efforts into sending adapters information
Currently, the adapter system is simple, i.e. it outputs information. One goal is to eventually be able to take external commands and run them on machines that are capable of doing so. While I believe this goal is far from being achieved, the most basic form of this is to be able to take information and process it within the adapter (Without interfacing with physical equipment).
One simple example of this could be the ability to manipulate the internal state of the adapter. A use case could be an external tool monitoring experiments (or the equipment) and trying to analyse if errors have occurred. If the tool thinks errors have occurred, messages are sent to the adapter, which could take action (changing its internal state, for example).
I think before this is developed further, it needs to be thought out more.
- What do we actually want to do? Is the example above really realistic? Could be actually be possible to detect faults?
- What would the adapter actually be able to do if an external tool detects an error? Send an experiment stop mqtt output? Detecting the end of an experiment automatically is hard.
- Security. Sending spurious commands to random equipment. We spoke, and this IS an issue. But like we said, without serious development, this isn't an easy solution. Therefore, it's up to us to agree on the assumption that if somebody has access to the mqtt broker, this is enough security.