earlier replies
Replying to @clemensv
You do get that not trusting and paying someone to operate that DB is the whole point, right?
1
I understand that being the assumption. I'm entirely unconvinced that it works out economically.
1
with POW, the total system cost of a blockchain tx are astronomically higher than putting a record into 2+ DBs
3
1
Additionally, sometimes even setting up the business relationship with the intermediary is either impossible or way too costly
1
Here, the focus is one car and you want proof of monotonically increasing readings from that car. It's not even the right structure
1
How so? If over a 10 year car history, I get 20 entries that increase monotonically, that’s sufficient?
1
you and I know people who drive 40k+ in a year, and sometimes 5k+ in two weeks. What if I reset by 500k once week?
1
Fair enough, can’t comment on whether that example is relevant, but sure, you need to record with the right granularity
1
if we're trying to prevent cheating we'll have to account for a hack that keeps setting the odometer back frequently in small steps
1
thus you'll probably have to track each change at the counter resolution of the odometer, i.e. each km.
1
Replying to @clemensv
My guess is the use case is to have a track record spanning multiple owners, only very few of which cheat.

Mar 18, 2017 · 10:10 AM UTC

1
Replying to @stilkov
for that, it's really enough to look into the manufacturer's servicing shop records. They already keep track. Globally, even.
1
But only if cars are serviced there. So one idea could be to set up a globally trusted intermediary … or use a blockchain :)