OK, first off - this is absolutely not on the cards, nobody is suggesting it (or writing the code afaik), but I got curious about this recently and just wondered what others thought.
How would people feel about a strictly anonymous, occasional upload, that could help the developers of AOG work out how best to improve the system? A one-way hashed unique ID for each installation, it could take the settings you use most, PWM etc, Pure-Pursuit, Stanley, etc and your average cross-track error, that sort of thing, and push to an online database. This would then give admins a way to trawl that data and go (for example) “well, nobody is really using Stanley” or “Integral is very rarely set at over 30”. Data like that might yield insights to help - for example - set more useful defaults out of the box or restrict the range of choices to guide users into making better decisions.
Just tossing the idea around, like I say, this is in no way in development or anything.
Also, if you’ve any suggestions on what sort of information you think could be helpful, post it below - likewise, if you’ve any concerns or similar, again, shout out
Vote is anonymous (as far as I can see) so have at it !
- Sounds useful !!
- No way !!
I think if it is disclosed upfront and opt-in most folks would be happy to help the development effort. Should also be able to opt-out at any time by changing the setting.
Absolutely, anything is possible.
The total anonymity would - I hope - allay most fears, and the fact that in isolation (ie, one person) the data would be useless. Only when aggregated could it give a decent picture. Thinking out loud, data would really only be of use to devs but I guess there wouldn’t be any harm in sharing it in that aggregate form. And if we did go down that road, the data is sent could be saved in a log folder, in plain text, so people could see exactly what it contained.
Anyone got any suggestions for the kind of thing that might be useful?
I originally thought of this in terms of what PWM settings etc might be, Stanley, PP and how it tends to hold the line, that sort of thing.
Honestly not sure if anything meaningful could come from that, just chucking ideas around really. Are there too many options in AOG, a percentage of which are always within a couple of points of each other, does nobody really ever use the roll filter, and so on.
I think it would be very useful to collect this data. You could probably also find relationships between setting and could consolidate all of the adjustments into a few sliders. Most of the big commercial systems only have a few values that can be adjusted and they still steer pretty well. This would make setup a lot more user friendly.
I’m having issues with the boundary contour. It doesn’t throw errors would there still be away to track what’s going on with this?