Continuing our discussion from the last session, our next session will get into more depth on how errors for building infrastructure devices should be represented.
Some questions to consider:
- What level of detail needs to be standardized for error messages?
- Is it enough to simply communicate that the devices is unusable?
- Should the standardized error messages also provide enough information for a technician to troubleshoot the device?
- Should detailed troubleshooting information be provided through a separate non-standard channel instead?
- How efficient should error messages be?
- A simple error code is high performance and allows for millions of possible error types but then can only communicate the presence of one error at a time
- Bitsets could express multiple simultaneous errors with high performance but then the number of error types is severely limited
- Dynamic arrays of error codes can communicate many types of errors with no limit but then heap allocations are needed
- A string of serialized JSON could represent unlimited types of errors and provide troubleshooting information for them, but then heap allocation and string parsing are needed
- Should standardized error definitions be specific to each type of building device, or should the definitions be abstract enough to use across all/multiple devices?
- E.g. are doors and elevators different enough that they need their own error code definitions?
- What kind of errors should we expect to report for each different type of device?
We will be seeking input on all of the above questions and more. Please come armed with examples of your most hated device errors that you think a good standard should be able to express.