17 ms·
> They may blink fine for a awhile but then stops for a period, and then suddenly all missed blinks blinks in a rapid succession If you had a windows 95 PC con
by londons_explore 4y ago
> They may blink fine for a awhile but then stops for a period, and then suddenly all missed blinks blinks in a rapid succession
If you had a windows 95 PC controlling the blinker, the exact same bug and symptoms would be totally expected...
Do how in 25 years have we not managed to quash this class of bugs? (Where a simple task needs to be done at a certain time, and gets delayed).
- michaelmior 4y agoI don't imagine any class of bugs will ever completely disappear. Sure, we know how to write software that is extremely unlikely to suffer from certain classes of bugs and we have better tooling for testing, but ultimately it's dependent on humans.
- foobarbecue 4y agoIf timing is critical, you need an RTOS. (Real time operating system) Or just don't use software. I wouldn't expect the blinker delay to use software, just a timer IC.
- georgyo 4y agoIn older cars things like blinking lights was done with a dedicated circuit. By putting everything in software, we've actually made this class of bug appear more often.
- diydsp 4y agoThat's an underrated point: rust solves memory issues, but now schedule feasibility is outstanding.
- londons_explore 4y agoFor example, you could have a dedicated "do this thing at this time" processor. It can't be blocked, and if it accepts a job, then it is able to do it on time. The main processor then tells this processor to do all timed tasks.