There obviously is something unclear, yes. Having accurate mapping info and improving OSM are two separate things. I asked about the latter. If that data were actually crucial to Uber's business, what possible rationale could they have to share it for free?
But while we're on mapping data - I don't think "higher quality" data is in any way crucial to their business. (Owning the data so there's no dependency on third parties is another question). Uber's routing is accurate enough to work reliably.[1] What does higher quality buy them?
From where I stand, you're proposing that a company that's already bleeding money like it's going out of style should spend a ton of money, on building technology they have little experience with, to collect data that is of dubious value to them, to then give it away freely.
I get that you'd like OSM improved, but I really don't think Uber will be the one to do it.
[1] Based on purely anecdotal reports from my friends and my own experiences, the usual case for misrouting is Uber drivers not paying attention to the map, or thinking they know better than the map.
> If that data were actually crucial to Uber's business, what possible rationale could they have to share it for free?
I'd understand if this were 1987 and you were asking this question. But it's not, and even Microsoft is on the public collaboration bandwagon.
We begin by recognizing that if you want to get into mapping and not pay out the nose for it, then OSM is your starting point. From there, it follows that we defer to the lessons of the last two decades which show that when given the choice between either maintaining a private fork or upstreaming, then upstreaming is overwhelmingly the best thing to do, even from a position of pure self-interest.
I hardly think their competitive edge hinges on high quality mapping data at the exclusion of others. Accurate mapping data is an enabler for them to go faster, do better, for the things they are doing now. If that means drivers are making 3 deliveries per hour instead of 2, then the proposition holds true. Or if someone can be picked up/dropped off at the curb in front of their building, rather than the clubhouse bearing the street number for their sprawling apartment complex, then the proposition holds true.
> data that is of dubious value to them, to then give it away freely.
No. My position is precisely that they do have a business interest in this kind of data. I've said as much. This is the third time now.
And it's not as if Uber themselves don't already have similar initiatives to release their datasets. They've had a program for the last ~two years for sharing trip data, and two weeks ago they put a public face on their "Uber Movement" initiative to make it more widely available. And that's a dataset that's even more proprietary—there are tons of people in the mapping game, but almost nobody besides arguably Google has access to trip-level data at the scale that Uber has and is chosing to give away.
But while we're on mapping data - I don't think "higher quality" data is in any way crucial to their business. (Owning the data so there's no dependency on third parties is another question). Uber's routing is accurate enough to work reliably.[1] What does higher quality buy them?
From where I stand, you're proposing that a company that's already bleeding money like it's going out of style should spend a ton of money, on building technology they have little experience with, to collect data that is of dubious value to them, to then give it away freely.
I get that you'd like OSM improved, but I really don't think Uber will be the one to do it.
[1] Based on purely anecdotal reports from my friends and my own experiences, the usual case for misrouting is Uber drivers not paying attention to the map, or thinking they know better than the map.