I’ve spent quite some time looking at the alternatives, with the result that no matter how much I hate the specification, how horrible I think the format is, how terribly designed is the syntax, DBC is the standard that everyone seems to want to follow.
For our holy grain of DBC signals -> OVMS metrics, we can use ‘attributes’ to define OVMS specific stuff (like signal name to OVMS metric names).
So, I’ve created a DBC component, and put in the basic structures to define the objects in a DBC file. I’ve also started the work to parse DBC files and load them in. The basic idea is:
- There is a core MyDBC object.
- DBC files can be loaded into that object (either from VFS, or as static strings), unloaded, etc.
- The MyDBC object will contain functions to access DBC files, and maintain them.
- Functions will also be provided to export DBC files.
- Statically loaded (via strings) DBC objects can be overwritten and modified (especially during development).
- The vehicle base class will be extended to allow a DBC file to be associated with a particular CAN bus - that will then automatically handle metrics from DBC signals.
- RE TOOLS will be changed to use DBC files directly (in particular to use DBC modifiers, rather than the current unique ID approach, but also for things like DBC signal extraction and display).
The overall goal is:
- The ‘old’ (current) way of doing things is to write C++ code in a vehicle module to match CAN bus #1, ID 0x102, and treat byte #3 an unsigned integer to set it as the SOC metric.
- The DBC way is to instead create a DBC file describe CAN bus #2, message ID 0x102, and a signal at byte #3 (16 bits in, length 8 bits) mapped to the SOC metric. Then, just tell the vehicle base module where that DBC file is and let it handle the mapping automatically. No code. Just DBC description files.
- We can still have custom C++ code, for custom metrics.
- When reverse engineering, we can dynamically modify a loaded DBC and instantly see the metrics change. We can then export that DBC out, and include it in the firmware for others to use.
The biggest amount of work is the DBC file handling itself. The changes to the vehicle base class are pretty simple.
Regards, Mark.
The whole decoding/encoding thing (where we use DBC or whatever to say key X byte Y is converted to v.bat.soc using this formula) is of course the holy grain of this; and is built upon all the foundation I’m laying now. The end goal remains for vehicle modules with just a DBC (or whatever) and no actual code.