i believe we have discussed this at some point, and also related to #42 , but i can't find an issue that settles the issue.
so especially re: the #163 bug where bit-flips cause header data to just be wrong, i wonder how hard it would be to add some kind of parity data into our streams, balancing the overhead, and whether this would even be computationally feasible on the scopes, of course, which i have no idea about.
It would be great if (ideally, if the overhead was not too great) we had a per-line (i.e. line of pixels within a frame) parity check of some kind - either a single bit or something more robust, i haven't surveyed possible approaches. if not per-line, then per-buffer.
it would be good to have a "proactive" error detection mode in addition to the "retroactive" detection modes we have now, but opening this as a question because i'm not sure what's possible.
i believe we have discussed this at some point, and also related to #42 , but i can't find an issue that settles the issue.
so especially re: the #163 bug where bit-flips cause header data to just be wrong, i wonder how hard it would be to add some kind of parity data into our streams, balancing the overhead, and whether this would even be computationally feasible on the scopes, of course, which i have no idea about.
It would be great if (ideally, if the overhead was not too great) we had a per-line (i.e. line of pixels within a frame) parity check of some kind - either a single bit or something more robust, i haven't surveyed possible approaches. if not per-line, then per-buffer.
it would be good to have a "proactive" error detection mode in addition to the "retroactive" detection modes we have now, but opening this as a question because i'm not sure what's possible.