How Facebook, Apple and Microsoft Are Contributing to OpenStreetMap | Hacker News Reader

How Facebook, Apple and Microsoft Are Contributing to OpenStreetMap

theodi.org

392 pointsephesee175 comments

Seems like OSM has a future. This is great. Only sad thing is perhaps some day we won't be able to have the fun of making contributions ourselves, if the big companies handle all that for us like Google does. In the mean time, though, I can make an easier case to people as to why they should care about this, since it's now used by Apple and Pokémon Go.

OSM worked great for me on newly built streets in Cambridge UK, where Google Maps had zero cartography.

The only major disadvantage of OSM-derived applications is the lack of real-time traffic information. Google Maps harvests data from phones inside cars and some car navigation systems. They also buy it from third parties.

http://opentraffic.io/ is trying to do some data harvesting, but it's still not very good. Another option would be to build an aggregator. Lots of governments publish real-time traffic data, so it should be a matter of integrating that into an OSM layer.

Lastly, there's this old technology called Traffic Message Channel which uses RDS (digital data on FM radio) to broadcast traffic warnings. It's quite popular in Western Europe. There's already a F-Droid application that is trying to provide an interface for that: https://f-droid.org/en/packages/com.vonglasow.michael.qz/

I cannot stress how important live traffic information is. Without live traffic, navigating on densely populated areas is a nightmare. It's the only part of my mobile workflow I can't do using F-Droid applications. I'm currently forced to use a web version of Google Maps.

Here at Mapbox we're doing navigation on OSM with traffic data - we collect our own telemetry, and use it to guide OSM-based routing. There are numerous Android/iOS apps using our API that you could use, or you could build one yourself.

Please elaborate. From where and how do you get the traffic data from, and is it sufficient?

Our traffic data is generated in house using anonymized location telemetry from our SDKs, much like most other traffic providers. The live speed predictions we generate for congestion and ETA models are competitive globally, providing the most accurate ETAs available in many parts of the world. Our accuracy today is particularly strong in the US, when compared to similar services.

Do you know if Mapbox has any plans to work with the World Bank's OpenTraffic[0]?

[0]: http://opentraffic.io

Not at this time. It's an awesome project, but we started developing our traffic engines around the same time and needs quickly diverged. We do try to open source many of the low level pieces of our model as reusable libraries whenever possible, such as our graph normalization algorithms[1]. At the end of the day, much of the core is difficult to decouple from data engineering infrastructure which will be internal to each organization, due to the immense volumes of live data that must be handled to power a modern traffic engine.

Aspects of the model like noise reduction, modality classification, and speed distribution estimation also require lots of fine tuning that is specific to the telemetry being ingested & the usage patterns of the output. For example, our speed models learn and correct from observed errors in various situations over time, which is coupled to our internal metrics data.

[1] https://github.com/mapbox/graph-normalizer

I've been wondering, given that you've made all of the building blocks, why hasn't Mapbox built and released an open source maps app for Android? It looks like you're just weekend's work away from having an open source app which is competetive with Google Maps on F-Droid.

OsmAnd already exists, and it had a metric ton of development time put into it.

I second this, and very interested in doing it myself!

I presumed they open sourced it already since the "Maps" app in the fdroid store is the same app just minus some ads or something

The "Maps" app in fdroid[1] is actually a recompilation of Maps.me[2], which is from another company, not Mapbox.

[1] https://f-droid.org/en/packages/com.github.axet.maps/ [2] https://maps.me/

Mapbox has been pretty strategic about not being perceived to compete with it's developers. Besides creating an amazing stack remaining neutral I think is part of their appeal.

> OSM worked great for me on newly built streets in Cambridge UK, where Google Maps had zero cartography.

Cambridge is incredibly atypical, though, given its large number of smart people without much money.

New builds are a target of opportunity for local mappers. There only need to be one or two in a medium city to spot these. It's a much more _fun_ activity than, say, maintaining POIs.

> Only sad thing is perhaps some day we won't be able to have the fun of making contributions ourselves, if the big companies handle all that for us like Google does.

There will always be room for individual contributions, because there are things that the big corporations do not care about. For example, many hiking trails and agricultural tracks good for cycling are missing from Google Maps because they simply do not have a business case for those details. Yet they are on OpenStreetMap because when OSM editors actually walk or cycle those ways, they add them to OSM.

Another example is opening hours for shops. There are only three ways those can end up on the map: 1) the business itself adds them, 2) a large corporation is aware of them and adds them, or 3) an individual mapper copies them off the sign at the shop’s entrance. For a lot of small shops, only 3 is realistically going to happen.

Yeah, it also crossed my mind that once the basic stuff is in place, it frees us up to add more of the niche interest stuff.

I'm pretty sure that Google Maps does auto-discovery of hiking and biking trails by simply watching peoples' positions?

At least I've had rather good experiences using Maps for biking. Even small, unofficial paths were mapped, and travel times indicated that Maps had an accurate idea of their condition.

Google's pretty weak for hiking trails in the US in my experience; OSM is far from perfect but it's much better.

They're both significantly worse than maps focused on hikers.

I recently switch https://onthegomap.com from google to OSM for walking paths between points. The trail quality is much better where I live (Massachusetts) but I have been getting a lot of complaints from people all over the world that it no longer has their trails. I wonder what the lowest friction way to get my users to help contribute trails to OSM would be?

A built-in editor that's specifically targeted to what interests them.

There's an app called OpenVegeMap which shows you vegetarian and vegan restaurants nearby. It also allows you to add and edit this information. Found a nice vegan bistro? You're just two clicks away from adding a missing tag about it to OSM. Just this and nothing else.

Great idea, signing up for an OSM account is a pretty high barrier though. Is OSM is ok with anonymous edits through 3rd party apps?

No. The person doing the edits should be directly contactable through the OSM messaging system.

As Max Erickson says, they need their own OSM account. You can write your own editor, which submits changes to the OSM API. Users can authorise your editor to edit in their name via OAuth. Once someone has auth'ed your app, they can stay entirely within your app.

But be careful that you don't lead your editors to think this isn't a global database that others use, or other OSMers will be annoyed.,

You can act as proxy like WheelMap does (upload the edits on behalf of the users), but you must then be reachable when some edit is wrong.

StreetComplete is an Android app for answering simple questions about things in OSM.

https://wiki.openstreetmap.org/wiki/StreetComplete

Have an opt in for sharing their gps traces and maybe get them to take and share georeferenced pictures of trail signs.

With auto-discovery of routes, Google might be picking up the most popular routes, but that does that does not mean they are the best ones. For example, on Strava heatmaps you can frequently see routes that are heavily cycled, but they might also be extremely stressful because one has to share the road with traffic. In fact, most cyclists might only be using them because they are sports cyclists with very thin tires unable to explore and find other options. The great thing about OSM is that it only takes a single mapper to find a nearby alternative and add it to the map.

> At least I've had rather good experiences using Maps for biking.

You wouldn’t in many countries around the world. Google Maps is considerably inferior to OpenStreetMap for cycling in much of the Western Balkans, Africa, and Patagonia.

Funny you should mention Albania: It's exactly where I learned never to doubt Google Maps.

Specifically, never to assume that "It's like ten meters between these two points... I don't see a reason why we would need to go all the way around".

(There was a 30m deep gorge)

More to the point: Given enough position data, Google should be able to assess all those road parameters you mention, far better than manual entry could ever do at scale. They could easily notice, as but one example, when locals routinely take different routes than they would suggest and modify their data accordingly.

Shouldn't matter if Microsoft and Apple are both collecting the same telemetry and feeding it into OSM.

> (There was a 30m deep gorge)

which may well have been explicitly marked, tagged with depth information, in OSM.

> Another example is opening hours for shops.

Unfortunately, the dominance of Google Maps makes 1) more likely to happen for them - more likely even than 3) is going to happen for OSM. Or at least, that's what my gut feeling says.

A small-town or rural convenience store in most of the world is not even particularly aware of how urban shops (but actually not all that many, comparatively speaker) are adding themselves to Google Maps. So, 1) is indeed unlikely to happen. But because OSM editing tools for newbies, like StreetComplete, encourage people to add opening hours for all nearby businesses regardless of size, that information does make it to OSM.

In my own neck of the woods, if you are cycling around the countryside, you can not only find a nearby convenience store to stop for a drink or snack, you can even know if it is presently open. Meanwhile, Google Maps has almost none of the small shops in this region, let alone their opening hours.

Sure, I think in rural areas, 3) happens somewhat more for OSM than 1) happens for Google Maps. In urban areas, though, I'd guess that 1) happens far more often for Maps than 3) for OSM.

I'm also not quite sure how often large chains are importing their opening hours into OSM? Because quite a few of them probably do so for Google Maps (or at least - Google Maps often has their opening hours).

> I'm also not quite sure how often large chains are importing their opening hours into OSM

Very rarely, to almost not at all. Most OSM imports are still either roads, address points or buildings. PoI imports are very rare (to be almost non-existant)

Why wouldn’t it have a future? We use it quite a lot in the European public sector where we can’t use google maps because of privacy issues.

We’ve written a small app that does route calculations for us to make reporting business related driving (you get a refund, but that has to be taxed in specific ways) an automated process. We use OSM in that, but we do pay a company to do the routing algorithm and supply the data in a way that doesn’t abuse the standard OSM api. I’m not sure exactly how they do that, I think they host part of OSM or something like that. There is a big industry for small companies which are specialized in those things in Europe, and they are typically cheaper and easier to use than google.

We also rely heavily on OSM in our GIS department, where we have our own private maps, though they aren’t so much for traffic we have been building our own “street-view” thing, using OSM and our own 360 images.

The disadvantage of OSM is actually the submission process, like if a street is wrong, and you supply the correct information, it’s not always accepted right away. Which is a huge issue if you rely on it for something serious. That hasn’t been an issue for us for two years though, but we did have one week where an elected official couldn’t submit his driving because of a wrong one-way street.

> The disadvantage of OSM is actually the submission process, like if a street is wrong, and you supply the correct information, it’s not always accepted right away.

In what way does this happen? I haven't experienced it yet and I get the impression my edits are accepted automatically. Is someone reverting your changes with no good justification?

I had a change denied because I didn’t attach open source evidence, so it was my fault, but because it got contested I couldn’t just redo the change.

An account vetting system would solve this. Being an official municipality it’s kind of silly that we can change the actual direction of a street, but not use our internal maps as evidence for it. ;)

That shouldn't be a problem as long as you provide sources for a contested change. E.g., a photograph of the road signs, or simply explaining who you are in a comment on the changeset that reverted your edit and providing a screenshot of your internal maps.

Most local mappers will gladly work with you in resolving such issues. Try the country-specific subforum on https://forum.openstreetmap.org/ if you get stuck.

> if a street is wrong, and you supply the correct information, it’s not always accepted right away

I'm a long time OSMer, and I'm a little confused by this, because OSM is a wiki. There are no data controls. When a user makes a contribution, it's saved into the global OSM database straight away.

Can you explain more about what you've experienced?

> Only sad thing is perhaps some day we won't be able to have the fun of making contributions ourselves, if the big companies handle all that for us like Google

There is some opposition within the OSM community from getting big tech corps to do all the work.

Yahoo (remember them?) allowed OSMers to use their aerial imagery from 2007 till 2011. Microsoft has allowed OSMers to use Bing aerial since 2010, and to be honest, it's probably been the most beneficial contribution to OSM of the ones in the article.

Also, Apple gave a talk at SotM about "working with the community" and demanded that that talk not be recorded (unlike all other talks thre) or shared with anyone else. Hardly in keeping with the spirit of "openness".

Facebook's computer vision imports are of dubious use.

Indeed the Bing imagery is one of the greatest non-financial gifts ever given by a technology company, regardless of motives.

When I started working on OSM about a decade ago my area was a greenish blur on the photo underlay and everything had to be manually positioned from GPX tracks. It took hours to map a street and given GPS error margins it was never certain to be exact.

Now anyone can log-in and start tracing roads / paths / buildings and it just takes a couple of physical visits to backfill the enumeration and names.

Just to point out that the Apple talk was not recorded due to the company wide policy. (I was at SotM) On the FB import you still think of the first import, but after that they worked a lot with the local communities.

Feel free to install on Android: https://play.google.com/store/apps/details?id=com.mapswithme...

or https://play.google.com/store/apps/details?id=net.osmand.plu...

even better...don't need to use Google Play: https://f-droid.org/en/packages/net.osmand.plus/

also, the gp's first link (MAPS.ME) is available on FDroid as an open source fork with ads & blobs removed: https://f-droid.org/en/packages/com.github.axet.maps/

It also comes with some nice extra features added, but ymmv on the stability of those features

last time i checked you could not login with your account and update maps, is it fixed?

For anyone interested in setting up their own OSM tile server + Nominatim geocoding: Here's an up to date docker image that is relatively easy to set up: https://github.com/drwitter/osm-tiles-docker

Great for maps in offline environments and/or querying without API limits!

Really think OSM is great.

There is also some controversy around it : the big players do data mixing from different sources, which is not strictly compliant with OSM licenses. Also some say that their bulk editing creates nice map data, but harms local potential mapping communities, with early case of Netherlands where AND contributed map data resulted to no real manual map updates afterwards. So some say that community, not map must come first, so the “guerilla mapping” is bad.

> There is also some controversy around it

Do you have a link to some commentary on this?

I'm aware of MS's generated US building footprints[0] but this hasn't (yet) been integrated into OSM (largely I think due to quality and bulk-import etiquette worries). And also StreetSide[1] but this isn't imported into OSM itself, rather it's displayed as an aid for editors.

Facebook is looking at contributions similar to MS's US building footprints (but for roads), but again this is mainly around tooling and generation rather than data sources, and is speculative as far as I've seen.

[0] https://github.com/Microsoft/USBuildingFootprints

[1] https://github.com/openstreetmap/iD/pull/5050

> Do you have a link to some commentary on this?

https://wiki.openstreetmap.org/wiki/Talk:AI-Assisted_Road_Tr... https://forum.openstreetmap.org/viewtopic.php?id=57942 https://forum.openstreetmap.org/viewtopic.php?id=63456 https://forum.openstreetmap.org/viewtopic.php?id=57387

During a rotational program out of college I was the product manager at Facebook working on maps as the team led up to the first ML contributions to OpenStreetMap (https://2016.stateofthemap.us/how-can-ai-help-us-make-maps/). There's an interesting story about a large mistake I made that is interesting to share:

I started working on this project because I believe that map data should be open as did everyone else on the team. There was (and still is to some extent) a lot of skepticism about whether large corporations have the best interest of the map at heart - and what tools need to be put in place to make sure the quality stays high. It's pretty clear to me that there can be an alignment in goals because whether it's a company, person, or nonprofit everyone prefers accurate map data. The questions are about margins of error & OSM rightfully leans toward only extremely high confidence.

One good thing about Facebook joining in on updating the open map is resources are not much of an issue. Our team was able to just call up and purchase the rights to the satellite images we needed instantly. For quite some time we had working models that generated edits for roads that were missing from the map. Leading up to the yearly conference there was a question of whether we should present all of our findings to the community and see what they thought before making any edits.

No one on the team was willing to push the button on submitting any of our edits because we weren't sure if they were good enough and it would feel shitty to be the one who broke the relationship between Facebook & OSM. Beyond the software engineers even the people who were paid hourly to train the models became invested in believing it was important that Facebook reach a way to channel its massive resources into the project.

Eventually we all gathered up and I just pushed the button to contribute a random 1km / 1km square in Egypt (we'd already computed edits across the country accounting for over 100% increase but things could always get better). Then we waited. No one ever reached out so we started contributing a little more at a steady rate to see if anyone else working in the country noticed (including improving the original 1km / 1km box as the models improved).

We were testing the ML in some of the harshest conditions we could find (low contrast in the desert) so the research team started working on another model for India (winding dirt roads, lots of rivers, tree cover). All of our edits went through street-by-street human approval which was working but slow going since the number of people on the project could fit in a Facebook friend recommendation unit. After not receiving much contact in Egypt (the map community there is small) we received initial positive feedback from local mappers in India and decided to ramp up the speed of contributing which meant getting a war room in Menlo Park and rounding up a larger group of people to review the machine suggested roads.

I ended up gathering a few too many people and myself along with the core team was too hasty in communicating the quality bar for submission. We planned to shadow edits as the week went on to make sure the new members were up to speed but things unfolded much more quickly. Within a few hours we accidentally submitted a few bad roads. A local mapper noticed as he'd just driven the area on his motorcycle the day before. I immediately left a meeting where we were negotiating buying more satellite imagery and jumped on a bike back to the war room. No harm done, we had scripts ready to undo edits.

When I got into the war room things were more problematic. Quite a few of the new folks had made similar mistakes so we paused everything. There was now a small group of local mappers in an IRC channel worried about large scale vandalism (though they quickly realized that wasn't the case). They noticed the breadth of the edits and tracked down the accounts of most everyone in the room. The map community in India is one of the better communities but still a room of this many people making edits at a such a scale was unlike anything you'd normally expect since the ML made editing 10-100x faster than hand tracing imagery.

During the time when the local mappers believed this might be intentional vandalism they escalated to the global community. So that's how a large chunk of the OSM leadership found out Facebook was doing work on the map - some of our worst edits. Shit! It ended up not being so bad however. The entirety of the bad edits were reverted in a few minutes with oversight from the head of the OSM data working group. IMO the situation accidentally showed some of the OSM corporate contribution skeptics that even the worst case wasn't so bad & Facebook was definitely wanting the community in control of any crises.

It wasn't all thought through but it's quite interesting in retrospect how the engagement strategy panned out.

Hi from the OSM data working group. Please don't do massive imports like this without discussing them first with the local community, and the global one if there's no significant local one.

The impact from the massive Facebook edit was more important, I think, and people are still very skeptical about whether massive ML mapping like this can provide value to OSM today. Announcing and discussing the edits first would have been better PR. Recovering goodwill will take a lot of time and effort.

Our guidelines prohibit undiscussed edits like these, and not only the ones that cause trouble. It is not respectful to the unpaid volunteers that cleaned up after this that LukeWalsh seems to think it was "not […] so bad".

Please see:

https://wiki.openstreetmap.org/wiki/Automated_Edits_code_of_...

https://wiki.openstreetmap.org/wiki/Import/Guidelines

Uber https://github.com/Uber-OSM/NewZealand plus many other smaller organisations have dedicated OSM teams contributing too.

Are Apple and Microsoft's map services built on OpenStreetMap? Or are they just contributing for the heck of it?

Yes, they both are.

Not originally, though Microsoft have recently added an OSM layer to Bing, and Apple have attributed OSM as being a contributor to their overall AppleMaps dataset in some parts of the world.

Apple Maps lists a whole bunch of data providers. OpenStreetMap is among them.

Maps are not just roads. POIS, transit rails, pedi paths all play a role,esp in cities. And OSM does well with path data but it's POI data is disappointing compared to Google. Transit stations are in the data but no info about what lines serve them. For eateries or stores, hours, menu focus, parking data is often part of Google POIs but is missing from OSM.

All great reasons this data is not present, but that does limit OSM usage in a wide group of cases.

The data that is present, however, is wonderful.

One of the challenges there is POI freshness. A paying advertiser is going to be timely across the Googleverse.

I do occasionally add opening hours to OSM but the interface is a bit clunky in id editor.

I would recommend using Street Complete, that has a really easy interface to add that kind of information to osm.

https://play.google.com/store/apps/details?id=de.westnordost...

It is in beta but still very useful, and does not feel like a beta (really stable)

This is great! Google has recently spiked fees drastically and it will be good to see more alternatives with high quality data

Because of that I needed to abandon Google Maps Api. But good thing is I learned how to download OSM data, generate and host my own maps.

Yeah, me too :-) Just got it running yesterday on my devbox. Still thinking of hireing Geofabrik to set up the real server when I'm done though.

This is great, but I'm surprised to see Apple featured so prominently in the title here alongside Facebook and Microsoft.

Apple has an internal volunteer programmer whereby employees can choose to spend their time contributing to OSM—part of employee benefits (which is nice), whereas Microsoft and Facebook are actually contributing mapping data the company has created/collected to OSM.

Furthermore, just 2 of the 3 listed companies are major sponsors of OSM events: https://wiki.openstreetmap.org/wiki/State_of_the_Map_2018#Sp...

Apple now employee their own teams to do mapping

Do you have a source? I'd be interested to look at the work they're doing.

https://github.com/osmlab/appledata

Thanks for that. I see now that it's linked from the article—I'd previously overlooked it as the article seemed to imply it was the product of their volunteer programme, but it seems it isn't:

> The building information was developed with sensors from low altitude aerial imagery.

> The building footprint and height information was captured and developed in 2013.

There's also Atlas[0], not mentioned in the article, which seems to be a (open-source/non-free??[1]) tool released by Apple for use specifically with OSM data. Really interesting stuff.

[0] https://github.com/osmlab/atlas

[1] https://github.com/osmlab/atlas/blob/dev/LICENSE

I got the link from my notes of the StateOfTheMap conference where an Apple representative was speaking. They have full-time staff now, claim 1 month training. It sounded like a dozen people.

> just 2 of the 3 listed companies are major sponsors of OSM events

Don't assume that because a company isn't listed, it doesn't/hasn't contributed.

> Don't assume that because a company isn't listed, it doesn't/hasn't contributed.

That is the list of all sponsors for that particular event. Apple did speak at the event, though their talk[0] was the only one not recorded, photos were forbidden, and they uploaded no slides afterward. The topic was their employees' volunteerism.

[0] https://2018.stateofthemap.org/2018/T081-Working_with_the_Co...

Yes, I was there. You might remember from the Q&As afterwards that Apple is a corporate member of OSMF but not listed as such, for example. The topic wasn’t principally volunteering, though that was mentioned.

The missing maps (https://www.missingmaps.org/) project is very popular in CSR programs around the world even if it's a bit "mechanical turk" for tech crowds.

Even google employees have contributed to the project by running internal mapathons in conjunction with organisations such as Médecins Sans Frontières...

Will be running some events in my current company in the coming weeks which will be fun!

So OsmAnd and maps.me look great. Is there an Android app which i could run to contribute traffic data flow and traffic events (accidents or obstacles on the road) like Waze does?

There might be 3rd party apps that do it, but OpenStreetMap doesn't store or support traffic data flow.

Maybe http://opentraffic.io/ ?

Someone mentioned it in the thread but I can't find who. There is https://openstreetcam.org that I haven't yet used but intend to.

Is there anything OSM-based like gmaps with the ability to search for ie restaurants in an area, then get voice navigation to the location? Even something that lets you search for "McDonalds near this address" would probably be good enough for me to switch.

Sure, but how useful they are will depend on whether the data in your area of interest is decent or not.

OSMAnd and Maps.Me both have poi search and voice navigation.

They could use some money and hardware... Currently the site loads

This website is under heavy load (queue full) We're sorry, too many people are accessing this website at the same time. We're working on this problem. Please try again later.

It is quite strange to see that many companies compiling essentially the same dataset for themselves. If a proprietary map is not necessary for your strategy it may well make more sense to go with OpenStreetMap.

I wish they contributed to DuckDuckGo as well. Or perhaps the issue is that DuckDuckGo is not open-source?

Microsoft provides the search index for DDG in the form of Bing.

Interesting. I wonder in what format the index is provided. (Hopefully then don't simply anonymize and forward user-searches to Bing).

I want to try openstreetmap, whats the best client available for android? (Pref. f-droid)

I use OsmAnd, it works great for me.

MAPS.ME is the best one by a pretty fat margin, I'd say.

Map viewing is wonderful, the best I have experienced on Android. It's very fast and super detailed.

Routing has some issues. The estimates are way off, routing by bike means road bike (not MTB) and the voiceover speaks too early or too late.

I've had a pretty good experience with Galileo (https://galileo-app.com/) on iOS; it looks like there's an Android version that's fairly similar.

I use both osmAnd and Maps.me from the fdroid store.

Maps.me: https://f-droid.org/app/com.github.axet.maps

osmAnd has a ton of features. maps is more basic but is better for navigation since it has a 3d option

OSMAnd is rather powerful, and its UI is not intuitive at all before you spend some time with it, but it's what I've come to like.

This is competitive pressures at work, similarly to IBM and other's investment in Linux in the early 2000's as a way to place competitive pressure on MS/Windows.

If you're not the leader in the field, and won't be able to catch up just by investing, you try to make the leadership in the field (Google's) irrelevant.

If the 2nd-best-by-small-margin base map info is open, having the best one isn't a competitive advantage anymore.

Google recently hiked enterprise license fees for access to translation services pretty substantially. I wonder if we'll see similar competitive momentum in that market.

Most likely, yes. But translation services are far easier to enter than mapping.

Both are hard and are open problems.

Sure, but very different. Mapping requires real-world operations, such as driving cars and taking pictures around the entire world, and doing it again and again.

This is far more cost-intensive, both in upfront as in maintenance, than the translation problem.

Software is easy. The real world is hard.

Which is exactly the point. Translation is a real-world problem. You need to actively go into the real world, find problem cases, make your translation software translate it, make your own translation, compare the translations, teach your software to translate it correctly, and hope that your approach is good enough.

The real-world constraint of translation software is building a sizeable quality corpus of real-world text with real-world living human translators. We're lucky that the EU employs literally thousands of translators to publish identical texts across languages (for free!!), which has helped make European languages pretty much a solved problem in machine translation. But anything non-european is still hopelessly incomprehensible, simply because we don't have enough real-world data.

Since they are using this strategy, translation of idioms is terrible. I have switched translate.

To what?

Far easier to enter, because so far there isn't a single one that is more than vaguely useful. As far as I'm aware, there isn't really much business at all for translation because at the end of the day, if you want it done anywhere close to correctly, you have to hire someone that actually knows both languages.

Second best? At least in my experience, OSM data is leaps and bounds better than Google Maps in every way possible, and it has been that way for a couple years now. The reason that I still use Google maps is that it's fast and well-integrated with my phone, and it has a lot of added features on top of the actual mapping that makes it very useful --- things like location sharing when you're on a longer road trip, creating a notification on your phone to route to an address from your browser, live traffic information, road closures, and business reviews. In terms of the actual data, though, the details available on OSM puts Google's data to shame.

I've been trying to get Google to fix an incorrect path drawn at a state park that I frequent. I've submitted an alternative route twice now, but it's been denied twice with little explanation. The path is a loop, but Google has it missing half of the loop, off by an eighth of a mile or so, and has it drawn where it cuts through private property. In contrast, OSM's data is completely spot on with what my GPS shows.

The park is in a tourist area, and I go to this park often enough that I've actually ran into multiple people visiting who have been standing on the trail with Google Maps open on their phone, thinking they've made a wrong turn.

I've experienced the same thing at GNP in Montana, BWCA in Minnesota/Canada, and all along the Richardson Highway in Alaska. If you're in a rural area or on public land, OSM data is the best available --- even better than what a ranger station would sell you.

> Second best? At least in my experience, OSM data is leaps and bounds better than Google Maps in every way possible, and it has been that way for a couple years now.

I'm sure that, for whatever product in the world, you can find a subset of people in the world who will swear that such product is, in they experience, the best in the world.

Mapping is not a subjective topic, though. I'm saying that OSM data is objectively better for the rural parts of the world I've been to.

An interesting benefit of Google Maps which (I don’t think?) is available in OSM: different maps for countries with boundary disputes.

Google is basically the de facto map, even as far as many governments are concerned, and so showing localized maps of disputed areas may actually be saving lives.

Osm tracks disputed borders. A vector tile based view could show differences based on location.

The tiles served from osm.org don't.

Yes, OSM.org shows one map for everyone. But you can take the OSM database, and make your own map to show whatever you want.

For example, the Indian OSM community have an OSM map, which shows the borders of India as per Indian law (w.r.t. Kashmir)

https://openstreetmap.in/#5/22.150/79.081 vs https://www.openstreetmap.org/#map=5/22.167/79.102

>for the rural parts of the world

I'm sure that's true. Google Maps seems to be better or at least more consistent (and better at routing with real-time information obviously) for the things that Google cares about. Which doesn't include rural areas, hiking trails, etc.

> I'm saying that OSM data is objectively better for the rural parts of the world I've been to.

Which is a subjective factor. "X is objectively better for the categories that I subjectively selected".

Mapping is a subjective topic, simply because there are endless ways to evaluate and compare maps, so attributing any type of weight to any factor is a subjective exercise based on your own personal preferences.

What's your reason for thinking Google is the best then? Do you even use maps for anything except major highways?

I'll agree that what layers a map ought to show is a subjective matter.

But if you have two maps and they both have a shared subset of layers, you can (for that subset) objectively evaluate which map more accurately represents the real world, eg: has more objects than the other map, and/or has objects marked more accurately.

And what I'm saying is that for the layers that they have in common, I think OSM wins hands down.

The place where you live and the places you happen to visit is a subjective factor. Other people (who visit other places) might have a different overall experience, even if you'd agree with them on any place you both visited.

It is a subjective topic. Each service provides different layers of data, with varying amounts of coverage for each layer. People have opinions on how important each layer is, and what's acceptable in terms of coverage for those layers.

Where I went to college, near Rochester NY, housing data is nearly non-existent in OSM, while Google Maps happily provides it and 3D models of them too. Where I am in Seattle, OSM has parking information while Google Maps does not.

Interesting, when I took Geography 101 the first thing they taught was “mapping is super subjective”.

Except for Lotus Notes.

Not too fast: http://www.ilovelotusnotes.net/

I only use google maps because the goecoding is a lot and I think the routing is done on their computers which makes it a lot faster.

Plus finding good OSM apps is hard.

Maps.me, Karta, Osmand.

> Karta

Contains ads, in app purchases

Hmm...

What are your objections to those?

I have bad news for you about Google products.

I really like Maps.me, however the main issue for using OSM is search. It is quite slow and it does not handle all address formats well.

I think OSM is a great data source and I always try to contribute at least the business opening hours but it deserves a better application layer on top of it.

Google's business location coverage in suburban areas is far better than OSM from my XP...by a longshot

Mileage varies depending on where you live. I really like the idea of OSM and have tried it on and off for years but it's abysmal for my city. Post code searches come up wrong and my own home address places the map marker the other side of town from my house. Google Maps has been accurate every time I've used it and I've never had any issues. I can't really trust OSM as an accurate source over Google Maps right now but hopefully it improves as time goes on. I really hope it does. I have tried to use open source solutions a lot in my life but I can't with Maps right now.

It's open for you to make the changes nessessary. It's only as good as it's editors and there isn't that many that can dedicate a lot of time to mapping. Thank you all that does!

This actually does work. I went though a phase of keen mapping and entered detailed information for my local area, especially for things off road. (Triggered by OSM becoming rubbish when the license changed most of the Australian data was removed.) Consequently for the last 6 years OSM is my fist stop for anything to do with my local area. Hopefully others have seen a benefit too.

It's always interesting to see where my OSM data pops up, as I mapped a number of obscure cliff lines in in accessible bush, which are quite a distinctive feature. For example, participating in rogaining or orienteering events and realising that I drew the base map 4 years before!

Wait, what happened to Australian data?

For anyone else wondering: https://wiki.openstreetmap.org/wiki/WikiProject_Australia#OI... and https://wiki.osmfoundation.org/wiki/Licence/Historic/We_Are_...

My problem with OSM, besides the lack of satellite layer, is that it doesn't seem to offer any (detailed) info of landmarks - just the name, sometimes not even that: for example, some churches are just marked with a cross.

If you want to try something really fun, install OSMAnd and download the Wikipedia maps, and then enable them I think in the "layers" option. I don't even think that's OSM data, I think it's geocoded Wikipedia articles. They show up on your map, and you can click on them and read them all offline.

I was way wrong. It's under "Configure map" and then "POI Overlay...".

If a church has its own Wikipedia lemma you can tag it. Most cathedrals have those. An image can be linked to, and of course the church website can be tagged as well.

For example:

    https://www.openstreetmap.org/way/4532022

But in areas with fewer mappers tagging will be limited:

    https://www.openstreetmap.org/way/390264609

It all depends on the local mappers adding and validating data.

How do you even get to this info page? I mean the "way" one.

I tried to click/double-click the text "Kölner Dom" (which is how it works on GMaps) on https://www.openstreetmap.org/#map=19/50.94132/6.95838 but it doesn't show anything.

On https://www.openstreetmap.org you mean?

By using the 'query features' tool on the right (the pointer with the question mark), and then picking the object you want from the list of objects in that area.

Other viewers of OSM data (like smartphone apps) can have different (sometimes easier) means of accessing that data. OSM is a database intended for use with many clients; https://www.openstreetmap.org is just one.

Thanks!

OSM is a very flexible data model. It's possible to add a lot of detail to things. Just because there isn't the data there for everywhere, doesn't mean it can't be added.

Google Maps beats OSM with absolutely no content in regards to POI data and navigation.

Especially navigation. Google can plan my trip from start to end, perfectly. There's not a single OSM-based thing that comes even close.

Galileo works pretty well for me, pathfinding and all.

https://galileo-app.com/

The alternatives suggested in this thread are all like: oh, X does A, Y does B, Z does C, etc.

But people mostly don't want three or four different mapping apps to cover their use cases. Having it in one app is a better user experience.

Doesn't have a web user interface, I can't plan a trip from my small phone screen.

I'm a fan of OSM. Where I am now (New Hampshire), Google beats the heck out of OSM if I'm searching for establishments.

(It's okay though, I'm happy to work toward changing that.)

In my limited experience with trying to use OSM for geocoding using Nominatim, it didn't work very well for New Zealand locations at all. It would generally map the GPS coordinates as being in the centre of the block, rather than the actual address.

Just another reply saying "OSM is completely useless where I live".

If I enter my street address into Google Maps, it shows my house. If I enter it into OSM it says "No results found" and can't even figure out what country or city I am in.

At least you were given the chance of improving it. You declined to do so, but that's your decision.

In a similar situation with Google Maps, you wouldn't even have been given the opportunity.

You can add missing places (including private property) to Google Maps, but I've never had an issue where I needed to do that.

Because everyone implicitly forced for improving the maps via a bullsh*t named captcha.

That’s not an extremely compelling argument. I could host a web-based persistent infinite bitmap canvas where anyone can draw anything they want, and that would still be a much worse map than Google Maps despite the fact that the friction to contribute is as low as it could be.

> You declined to do so, but that's your decision.

This kind of patronizing attitude is unnecessary and silly. The vast majority of users of any software product are only interested in being just that: users. And that's totally okay. No need for passive-aggressive shaming.

The vast majority who make use of OSM data probably aren't even aware of what Open Street Map is, nor that they can contribute data to such a thing, or of how they could do so.

After re-reading my own comment, I agree and am sorry. I meant to put the emphasis on "your decision", in order to highlight the freedom argument for OSM. I realise it totally sounded like the emphasis was on "you declined", which makes it a personal attack, and that was never intended nor necessary.

Again, sorry.

Good on you, you're a better commenter-person than I am.

Just out of curiosity, have you looked into what data is missing? Is the house (or at least the street) mapped, but missing the street name or the house number (assuming you live somewhere where these are relevant), or is the street completely missing?

OSM doesn't support/understand the street numbering scheme used in my country. If I rewrite my address in the format OSM thinks I should be using, then at least it can find the street I am on. But I'm not going to write addresses differently just for one website...especially when Google Maps understands how we write addresses.

Hi, I'm a long time OSMer, and I'd like to help. OSM has a flexible tagging system, so it could be possible to map your addresses the way they should be. You shouldn't have to shoe-horn things into a schema that doesn't fit you.

Which country are you in?

I am in Vietnam. Here are the problems:

Addresses here are written like "220/9 xo viet nghe tinh", which means "go to 220 on Xo Viet Nghe Tinh and turn into the alley there, go to #9 in the alley".

Typing "220/9 xo viet nghe tinh" into OSM results in "No results found". Instead I have to write "hem 220/9 xo viet nghe tinh" for it to find the road; no one writes it that way.

A house address would be "220/9/14 xo viet nghe tinh" meaning "go to 220 on Xo Viet Nghe Tinh and turn into the alley there; go to #9 and turn into the next alley you find there; then go to house #14".

The only way OSM can figure things out is if I write it like "14 hem 220/9 xo viet nghe tinh"

You mean the search box doesn't handle it well?

How does https://photon.komoot.de/ do?

Yes, it works. Well, it is missing the location but it at least understands the format. What is it doing differently from the OSM site?

It's different software.

openstreetmap.org uses Nominatim:

https://github.com/openstreetmap/Nominatim

The photon link uses Photon:

https://github.com/komoot/photon

Nominatim processes OSM data, associating items of interest with features like town boundaries that enclose the item, or place markers that are nearby, and then is also a fairly pedantic search interface for that data.

Photon takes the data that Nominatim generates and sticks a more flexible search on top of it.

Checking for/creating an issue for Nominatim describing the address format that doesn't work well would be a useful thing to do, as global support is a goal there.

As others said, it depends. Here in my city (Lübeck) in Germany, OSM runs circles around GM and then leisurely sprints towards the goal. When I was in South Africa (Pretoria) OSM was so out of date that it was unusable, and that's without information there that's always been wrong.

For anyone interested in improvement of OSM data, I'd like to recommend the StreetComplete android app:

https://play.google.com/store/apps/details?id=de.westnordost...

While contributing you almost feel like playing a game :)

Yeah, I used it for a while (lived there for half a year), but the base data was just too wrong for it to be much fun.

What's a good phone app to use it with on Andorid? (As I understand it, Open Steet Map is the data).

OsmAnd! It can download vector maps for countries at a time so you can navigate offline. It's a bit janky sometimes but hey.

CityMapper uses OpenStreetMap data for the cycle and walking route planning (even though the GUI is Google Maps). Best cycling directions you can get, full stop. The algorithm is by CycleStreets, a UK academic start-up out of a university.

Maps.me is more user friendly than OsmAnd and feels more like Google Maps, but needs internet and just feels a bit strange to me. Ad supported but you can turn the ads off in the settings.

And for iPhone?

I once reported an issue with a stadium in my city. I bike in front of it twice a day. The issue was that it was licated on the wrong side of the road.

They said that they checked and it is ok. I asked them how they did this because, well, I actually see the place twice a day. They said that they have ways to check (or smthg like that).

I then sent M their own stree view link and screenshot. Not good enough. So I have up, I do not need the gps to bike there twice a day.

Two years later, suddenly it was approved.

All my corrections are now approved automatically (within an hour or so) so they may have improved.

I also told them to check on OSM where I also updated some information about the surroundings :)

Google does not have the best base map info. OSM may well be the best global base map, actually. Google's advantage is their applications that use the data.

Citation please? What do base your statement of fact on?

OSM is generally richer in term of what can be documented. For example, you can describe a bench, how many seat it has, what color is it, is it in metal, wood or stone, and so on. This is not necessarily displayed on osm.org, but could be on a more specialized app using osm dataset.

For example compare the two following areas :

https://www.openstreetmap.org/#map=17/41.88846/12.49110

https://www.google.com/maps/@41.8884227,12.4916715,17.54z

Google Maps is lacking some details inside the park, you don't have any path, fence, grass, water tap, bench or bin.

The issue for OSM is that some less popular areas are totally barren.

Also note that the osm.org bitmap tiles do not, in general, contain all information OSM has about an area. Things like future roads, hedges, etc. may not be shown.

Not that I don't believe in your claim (I really prefer to live in world where OSM is better) but isn't how you form your argument is a kind of confirmation bias?

What are some OSM-based mobile apps that I can compare against Google with?

In what area? OSM data cover lots of use cases. For navigation on Android check out Maps.me or Osmand.

Our HR department wouldn't let me give proper feedback. Even though I've had some people that, with a little tweak to their knowledge, or a slight shift in thinking, could become great hires.

There's tonnes of candidates out there that just lack something small, or are slightly less ideal than someone else... yet still have all the raw talent and potential in the world.

But you can't always hire them because you don't have the resources to up-skill them efficiently. Or you need someone that can more readily contribute to a project.

I've certainly got some people in mind whom I'd love to contact as soon as I start at a different employer, to drag them on board if said employer had the resources to properly develop them.

Ah well.

Wrong thread, perhaps.

That's embarrassing.

Reply on news.ycombinator.com