6 ms·
Maybe the 400 years referes to the number of people-years used to manually clean up map information?
by ericHosick 14y ago
Maybe the 400 years referes to the number of people-years used to manually clean up map information?
- masklinn 14y agoThats how I read it. Although frankly speaking, if anything 400 man-years seems very low for worldwide mapping.
- dvanduzer 14y agoDon't underestimate the value of legwork and algorithms. edit: I... I skimmed the article the first time, but... The author actually claims that it's a data quality problem, not an algorithm problem. Huh?! second edit: I guess my only real contention is that "data quality" is a fake idea.
- dvanduzer 14y agoSo it's a problem that $4 billion could solve in a few days, right?
- blrgeek 14y agoIf you can get 14600 workers who can do this work and the software, you could do it in 10 days.
- dvanduzer 14y agoI thought it was reasonably entertaining to consider 400 salary years at $100k each just to see what the numbers looked like. Apple has had trouble with Big Data, in the parlance, but iCloud is still running. They've made some pretty dumb mistakes with their online services over the years, but they still do happen to be syncing data (if even just push notifications) to something on the order of a hundred million devices. Catching up to where Google Maps is today will take somewhere between 10 days and 400 years. I'll bet good money that they have this nailed within a 95% tolerance before next year's new hardware model.
- blrgeek 14y agoFrom what I've heard, Google had an army of around 1000+ contract workers in Hyderabad, India, cleaning up the data for about 5 years. If anything 400 man-years sounds low :) This is apart from MapMaker, and their dev teams, and their software improvements. For instance, if you look at directions in India now, they're awesome, since they tell you landmarks to look for on the way [pass by * on left], at the intersections, etc. http://goo.gl/maps/b8w15 http://goo.gl/maps/b8w15 for instance. My guess is 95% will take 3-4 years - once they accept they have a problem.
- dvanduzer 14y agoI don't think 3-4 years is a crazy estimate. This is pretty cutting edge stuff. If you hadn't already seen it, the Atlantic piece on the technology is a fantastic read: http://www.theatlantic.com/technology/archive/2012/09/how-google-builds-its-maps-and-what-it-means-for-the-future-of-everything/261913/ http://www.theatlantic.com/technology/archive/2012/09/how-go... On the other hand, this actually is one of the rarer problems where throwing 10,000 contract workers from India for 5 months could solve much of the problem, especially since Google has already paved the way on the harder parts. What I really want to see is a serious quality comparison between Apple and OSM. edit: And the million dollar question... How much of Google's map data was correlated/verified against iPhone positioning data?
- maxerickson 14y agoDoes the position data provide any information beyond "phone was within a meter or two"? I can see how that would help reveal issues, but I don't see how it would help much in fixing them.
- zalew 14y ago9 women won't give birth in a month.
- dvanduzer 14y agoTen developers each giving their boss a copy of Mythical Man Month won't make him read it any faster either.
- epo 14y agoHa ha! One of the better variations on this oft misquoted old saying I have read.
- shawndumas 14y agoNo but then you'll have nine babies in nine months.
- batista 14y agoGestation time is not a "parallelizable" problem. Fixing map data is.