3 ms·
More systems need to truncate location data. It is relatively "easy" to solve this particular issue by just not returning highly specific data. Instead of sayin
by ShakataGaNai 6y ago
More systems need to truncate location data. It is relatively "easy" to solve this particular issue by just not returning highly specific data. Instead of saying "896m", return "<1km" or "1km-3km" etc. Can you still abuse the data to some extent? Yes in theory, but it's significantly harder and takes a lot more work to get anything reasonable.
You can also truncate the data you have on users. 2 digits of Decimal_degrees precision is +/- 1KM. If you're doing an Uber-like service, knowing where a user is down to a few feet makes a different (which side of the street). But if you're just concerned with "Hey people nearby you are using this app too" (which is all too common).. Do you really need to know/save/keep where a user is beyond 1 sq km? No, you don't.
Yes, Apple has a coarse data that does something like this, but take it that one step farther. Don't need the info? Don't collect it (your app doesn't need to send it) and don't store it.
- enw 6y ago> Instead of saying "896m", return "<1km" or "1km-3km" etc. Can you still abuse the data to some extent? Wouldn't eventually samples of arbitrary precision have their "range donuts" overlap heavily at the user's location? This could probably be automated to track in real-time someone's position.
- KingMachiavelli 6y agoYes, you just have to spoof locations near the boundaries of the quanta. If they are less than 1km then test locations until they are in the 1km-3km group but you will know they are actually 1km away. Repeat between 3 and a dozen times at different angles and you can still determine a location.