Any such test is likely to be highly subjective and therefore open to bias and abuse. A breath alcohol content of 0.008% might be fairly arbitrary in terms of risk, but it's objectively measurable. A cop can claim that I was driving erratically just because he doesn't like the look of me, a jury can interpret dashcam footage unfavourably based on their prejudices, but a correctly calibrated breathalyzer doesn't lie. Drawing an a line in the sand between "sober" and "too drunk to drive safely" based on a biological marker of alcohol intake is almost certainly the fairest means of enforcement. There is of course the broader issue of who the police choose to stop and test, but there's no obvious technological solution to that risk of bias.
Typical roadside walk the line test might not be all that objective, as it's judged by the eye of the officer, and some people may be better walking the line, while still having reduced reflexes and being a increased danger on the road.
Though, I'm not an expert and despite this feeling arbitrary, there may have been studies comparing effectiveness of various forms of testing on predicting the harm of DUI, or comapring breathalizer with the walk the line test, or touch your nose test or whatever.
I'm not for locking people up, but I also prefer safer roads. And what helps and what does not may not be all that obvious on the larger more aggregate level.
If the officer wants you to do a field sobriety test you tell them blood test or GTFO.
The values they use to punish you is completely arbitrary (probably varies from state to state) and we do not take any other factors related to the individual into account which is important if we are not so keen on ruining other people's lives. Look for behaviors that show there are issues with coordination regardless of substances. We can do that. We already do it. We do not need to punish based on arbitrary values. You could fail the test and yet have perfect coordination. See my issue? It gets even more problematic when you consider alcoholics who are actually better off drinking vs. not drinking when it comes to coordination and reflexes. Anyway, the emphasis, in my opinion, should be on behaviors that indicate loss of coordination, reflex issues, etc. not some arbitrary value. It works well with prevention, too, so there is no need to give up on that.
> Argument ad absudrum: Swerving may not lead to an accident either.
I agree, but depending on the situation, it could easily indicate that there is an issue regardless of the substance used.
Anyway, thank you for the calm-headed response.