FAQ: Difference between revisions

From OpenStreetMap Wiki
Jump to navigation Jump to search
(An obvious question.)
Line 223: Line 223:


We also use [[Yahoo! Aerial Imagery]].
We also use [[Yahoo! Aerial Imagery]].

== What makes a road belong to a city? ==

This is often asked by beginners.
There should be a closed way marking the extend of the city with a "place"- and a "name"- tag as well as a sindle node with a name and place to mark where to draw the city-name.
In the actual map such a shape does not exist for many cities, thus only the distance to the node that marks the city can be used in these cases. For exceptional cases an is_in can be used.

== what shall I do for roads that have multiple values for a tag? ==

The database can do this but most tools do not. So comma-separating the values is used. e.g. nat_ref="B500,B550" for a piece road that belongs to both, the B500 and the B550 at the same time.


=Rendering Questions=
=Rendering Questions=

Revision as of 20:10, 24 September 2007

Template:Language-FAQ




Frequently asked questions

About the project

Why are you making OpenStreetMap?

Geographical data (geo data) is not free in many parts of the world, including the United Kingdom and much of Europe. Generally these places have given the task of mapping to various government agencies who in return get to make money by selling the data back to you and me. The USA is the only major counter-example where due to copyright restrictions on the government, they have to give it away free.

If you live in one of these countries, then your taxes pay for the mapping and then you have to pay again to get a copy of it.

The data you get doesn't change much over time, in general. Roads don't tend to get bored and move around. You can usually navigate in many areas with any post-war or more recent maps.

The data you get contains lies, or easter eggs, to catch out anyone copying it. These Easter eggs take the form of fake or missing streets, or features like churches and schools that don't in fact exist. If you make a map using their data, they can say "ah-ha! Gotcha!" from looking if you also copied these fake pieces of map. The map may also just be incorrect because for example you bought it a year ago and a path has been dug up in your local park since, or someone just made a mistake.

If you accept all of this then you still can't do anything with the data but photocopy it. In lots of places that's illegal too if you go beyond your fair use (or fair dealing) rights. You can't correct a street name, or add the pub/bar over the road, or use the data in a computer program without paying a lot of money. More money than you probably have. What about sending it to a friend, enclosing it in an invitation or posting it on a notice board? A lot of these are less legal than you might think.

Advances in technology like cheap GPS units mean you can now create your own maps, in collaboration with others and have none of the restrictions outlined above. The ability to do so allows you to regain a little bit of the community you live in - if you can't map it you can't describe it.

Why don't you just use Google Maps/whoever for your data?

Short answer:

Because that data is copyrighted and owned by people like the Ordnance Survey. Google/whoever just license it. If we used it, we'd have to pay for it.

Long answer:

Most hackers around the world are familiar with the difference between "free as in beer" and "free as in speech". Google Maps are free as in beer, not as in speech.

If your project's mapping needs can be served simply by using the Google Maps API, all to the good. But that's not true of every project. We need a free dataset which will enable programmers, social activists, cartographers and the like to fulfil their plans without being limited either by Google's API or by their Terms of Service.

At this point, the usual rejoinder is "Why don't you just get people to click a point on a Google map, then record the latitude and longitude in the Openstreetmap database? That's free, isn't it?"

Unfortunately not. The data used in Google Maps is sourced from NAVTEQ and Tele Atlas, two big mapping companies. They, in turn, have obtained some of this data from national mapping agencies (such as the Ordnance Survey). Since they've made multi-million pound investments in gathering this data, these organisations are understandably protective of their copyright.

If you collect data from Google Maps in this way, you are creating a "derived work". Any such data retains the copyright conditions of the original. In practice, this means your data is subject to the licensing fees, and contractual restrictions, of these map providers. That's exactly what Openstreetmap is trying to avoid.

Please don't be misled by considerations of software copyright, or of Terms of Use. The Google Maps API can be incorporated into open source projects, sure. But this only governs how you use the software - it doesn't have any implications whatsoever for the data displayed by this API, which is still under copyright.

(It's not yet clear whether it's ok to create a derived work from aerial photography: some readings of UK law suggest that you can do this without 'inheriting' the copyright in the photography. A definitive ruling on this could open up new avenues for Openstreetmap and similar projects, but in the absence of such a ruling, we're continuing with the approach of sourcing our own, 100% independent data.)

Further reading:

  • Google Maps' Terms of Service. Note particularly the 'Map Information' section, and that:
    • "Geocoding data for map content in Google Local is provided under license by Navteq... and/or Tele Atlas... and subject to copyright protection and other intellectual property rights owned by or licensed to NAVTEQ, TANA and/or such other third parties."
    • "Also, you may not use Google Local in a manner which gives you or any other person access to mass downloads or bulk feeds of numerical latitude and longitude coordinates."
  • The Openstreetmap mailing list archives. You may want to search them for phrases like "derived works", and for a thread in October 2005 called "London locations".
  • For UK users, here's a useful PDF of copyright legislation.

How can I get involved?

There are lots of ways to contribute to the Openstreetmap project. If you have a GPS unit you can use it to collect data and use our online tools to add the data to our collection. If you don't have a GPS unit you can still help. Ways to help are listed on the Getting Involved page.

Is there a mailing list?

Yes! See our Contact page.

Is there a forum?

Yes! The Content of OSM's first (unofficial) forum has been moved to here. The forum is currently in its early stages, and therefore at present, most active OpenStreetMap users use the mailing lists.

Can I show my support by purchasing cool stuff?

Yes! See the Merchandise page for mouth-watering goodies.

How can a project like this create accurate maps?

By the very nature of the wiki-style process there is no guarantee of accuracy of any kind. Then again, few proprietary maps carry a guarantee of accuracy, either. In fact, some have artificially-introduced errors.

The essence of a wiki-style process is that all users have a stake in having accurate data. If one person puts in inaccurate data, maliciously or accidentally, the other 99.9% of people can check it, fix it, or get rid of it. The vast majority of good-intentioned participants can automatically correct for the few bad apples.

And as they say, your mileage may vary. The Wikipedia project has shown that a large amount of good quality data can be collected but it can be difficult to weed out the inevitable errors.

Currently there are no processes or mechanisms, such as recent change lists and watchlists, that can be used to easily monitor edits within OpenStreetMap. I'm sure some mediating processes will be developed as soon as they become necessary.

At the moment the best way to answer this question is to judge for yourself. One way is to pick an area that you know well and use the OpenStreetMap viewer to see how well the map data (if there is any, it's still early days yet) corresponds to your own knowledge.

Why aren't you using Open Geospatial Consortium (OGC) schemas and software for OpenStreetMap?

There are many existing tools which people often recommend for OpenStreetMap to use. The problem is, most of them are hard to use and maintain for a variety of reasons, and people are very reluctant to volunteer to help set them up and run them. We used to use MapServer for serving static versions of our maps, but unfortunately we found it to be unthreaded, slow and hard to extend - we replaced it with mapnik.

It's not that OpenStreetMap as a community is against OGC standards, but OpenStreetMap has been built iteratively using the simplest approach that could possibly generate useful maps. The focus is on street data and maintaining a 'wiki' approach to editing where all changes are logged and can be rolled back. Off the shelf tools don't support that in the way we'd like.

Help is needed on choosing which OGC tools and standards to use, and integrating them effectively into our existing systems. Please get in touch if you can spare the time and expertise to do this.

Also, see Why not GPX for a similar discussion about why GPX was rejected as a transport format.

Why don't you add OSM support to existing products and/or rather than writing your own route planning / navigator program?

RoadMap is a promising PDA and desktop navigation software, which is open source and uses TIGER for US data, and (as of 2006) vmap0 data for the rest of the world. The RoadMap trunk does not yet support autorouting, but has speech synthesis (for street names, etc). Ehud Shabtai's FreeMap fork of RoadMap which is in active development, has support for autorouting and has been ported to PocketPC and J2ME mobile phones in addition to the systems supported by RoadMap. It also has an interface which is more adapted to the small displays on mobile phones and PDAs.

Roadnav is another valid alternative, although contrary to RoadMap it offers no PocketPC support as of yet, and only supports TIGER data, although Digital Chart of the World (DCW) support is planned. It does support autorouting, however, and 3D views and aerial photos (wow!).

traveling-salesman is a rather new one. Mostly interesting for java-developers at this point.

The newest release of Roadnav has preliminary OSM support.

How do I link to a particular postcode on OSM from my own website?

You can link to the search page for a particular postcode:

http://www.openstreetmap.org/geocoder/search?next_controller=site&next_action=index&query[postcode]=AA1%201AA

You seem to have a lot of existing map data. Where did it come from?

...

Editing and other technical questions

I have GPS data, how can I use it to help OpenStreetMap?

You can upload your GPS tracklogs to OpenStreetMap, so that you and others can trace over them to draw maps. You'll need to be a registered OSM user before you do this. See tracks that others have uploaded.

Once you've done that, you can use the GPS as a guide to drawing roads and paths for OpenStreetMap. You can do this by:

  • edit your data online, using Potlatch, an easy-to-use Flash program that runs within your browser; or
  • edit at home using JOSM, which provides a powerful way to edit GPS data locally on your computer into sensible tracks before uploading them to the server (there are other editors, too);
  • and if you don't have a GPS you can still help, see Getting Involved.

What images and maps may I use to make maps from?

Only images that are free of copyright restrictions for derived works may be used as maps to give you hints where your data should belong to. You must not use printed paper maps that have copyright restrictions, even if you scanned them by yourself. You must not use any images from "free beer" sites as Google Maps to get any coordinates entered in OSM.

Images, that are made by the US government and thus are public domain may be used.

See also the page about out-of-copyright maps.

I tried to download my town/city/region but it doesn't work

Chances are the area you tried to download is too large, and the server probably timed out before getting the data to you. Try on a very small area first to make sure its working OK. If thats ok, then you are best to download the town in smaller segments. Presently, you cannot download an area large than 0.3 degrees in either dimension.

If you really want large areas of data, the best approach would be to download the planet.osm file, which is generated weekly. This file is basically a snapshot of the OSM database and contains all valid data.

I want to create a very long way, but I cannot download OSM data for that area

In order to be able to easily handle long roads, you should not make ONE long road out of it. You should rather split the road into several ways. As a rule of thumb, no way should be longer than 10–15 km. Typically, they will actually be much shorter.

Applications like route planners for example will be able to easily join the ways to one road again. This type of applications will need to postprocess the OSM data anyway.

For motorways for example, it makes sense to make a way from one exit to the next. Also, intersections of motorways should be the point where you split a road into ways.

My login doesn't work from JOSM

Note that there are two different logins for the OSM project, one is for this wiki only, and the other is for the website and API. You need to register on the www registration page to actually be able to work with OSM data. Please also note that you used to log in to the OSM website using your email address, not your Display Name, as opposed to the wiki (nowadays, either the email address or the display name should work for logging in).

Why should I not begin development on a new editor, one better/different/prettier than current editors?

There are already several very useful OSM editors in development, and confusing the space with "yet another" platform will only duplicate effort. Please consider contributing to one of the editor development efforts already in progress, such as JOSM, Merkaartor, or Potlatch.

Why can a given segment belong to more than one way?

Allowing this situation in the data potentially decreases the size of the OSM database, since one can re-use a given segment with its accompanying latitude and longitude data in more than one way.

This is particularly relevant in this situation: Assume one "logical" way represents a train track, another "logical" way represents a footpath, and another a road. All three of these ways might physically coincide in the real (geographic) world — for all practical purposes. While the three ways would obviously differ in their metadata (tags), sharing their underlying segments requires less space in the server's database.

Why don't we spread the load on the OpenStreetMap database across a number of servers?

Sharing the OSM server load via MySQL replication, or BitTorrent, or carrier pigeon is often recommended, but the idea has so far gained little momentum -- in terms of real development effort. If you have the system administration or coding expertise to implement such a distributed system, please do not hesitate to volunteer on the mailing list.

As a starting point, you may want to familiarise yourself with the existing software infrastructure, as described in the Development pages.

I have geo-referenced photography/shapefiles/waypoints for my area, how can I upload them?

We'd love your high resolution geo-referenced aerial photography/satellite images if they are free of copyright restrictions for derived works, or you are the copyright holder and can grant us the relevant rights. Unfortunately we don't have a clean and simple way to add extra imagery yet, but please get in touch via the mailing list and we'll let you know when we do.

The situation is the same for shape files and other data formats - we're definitely interested to hear from you, but we won't be able to add things immediately, yet.

Why is the data so inconsistent?

"OpenStreetMap is a free editable map of the whole world. It is made by people like you." Which means the database will always be subject to the whims, experimentation, and mistakes of the community; this is precisely OSM's strength since, among other things, it allows our data to quickly accommodate changes in the physical world.

I think I found a bug, what should I do?

If you find a problem with the map editing applet or the website, add it to the OpenStreetMap bug tracking database. We use trac, which uses your OSM username and password.

If you notice something incorrect/missing/wrong/unintelligible with the documentation you're reading right now, edit it! It's a wiki! For wiki advice and extensive documentation, see the mediawiki project, who created the software this wiki is running on.

My GPX file didn't upload properly

Your GPX should consist of trackpt's with valid timestamp. The ele(vation) tag is optional and will default to 0. Something like below is acceptable:

<?xml version="1.0"?>
<gpx
 version="1.0"
creator="GPSBabel - http://www.gpsbabel.org"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns="http://www.topografix.com/GPX/1/0"
xsi:schemaLocation="http://www.topografix.com/GPX/1/0
http://www.topografix.com/GPX/1/0/gpx.xsd">
<time>2005-11-07T14:00:09Z</time>
<trk>
  <name>ACTIVE LOG</name>
<trkseg>
<trkpt lat="52.564001083" lon="-1.826841831">
  <ele>115.976196</ele>
<time>2005-11-07T12:03:31Z</time>
</trkpt>
</trkseg>
</trk>
</gpx>

There are two things the importer won't do. First, it won't take in GPS points without timestamps as they're to be used to work out speed and so on. The other is that it doesn't import waypoints and your file consists only of waypoints.

The reason for this is that if you reset many GPS units or download map data to them, then you often get copyrighted data put in the GPX. The most famous example is that if you reset a Garmin GPS unit then it will put the locations of the Garmin offices around the world as waypoints on the unit.

My Tracks have more points on bends than on straight paths, and/or the points are very spaced apart

This is due to the gps device being set to record on Auto, which saves space by recording fewer points on straights. To Make the gps record more points, go to the area where track settings are available (On your GPS), and change the recording method to either 'Time', or 'Distance'.

Time: Time will record points every 'x' seconds/minutes/hours. This can be changed on some gps devices. Having this setting will use up the memory fast, but will increase the points considerably. The disadvantage of this method of saving is that when moving slowly with the GPS device the points will become densely packed.

Distance: Distance will record points every 'x' Meters/yards. This can be changed on some gps devices. Having this setting will use up the memory fast, although relative to your speed. The disadvantage of this method of saving is that when traveling on straight roads fast, (motorways/highways), unnecessary points will be recorded. Also if you are to be tracking a small area, if the 'x' value is too high, your route will be unclear.

Alternating between these two methods of saving is advised.

Wrapping: If the option is available to turn on, or off the wrapping function, then having it off is also advised. Having wrapping on means that when full, the gps device will make room to record the latest section of your route, by deleting the tail of your route. It will delete the tail, point by point, at the same rate as new points are created. The exception would be when you are logging track data to a data card in some Garmin models. The tail data will deleted from the device's built in memory, but not the data card.

I get bad signal / traces in city centers / near big metal buildings

For a GPS to work and achieve some accuracy then it needs to receive at least three satellite signals. Often when in city centers there are many tall buildings that can block these signals and stop the GPS from being able to work out where it is. There may also be multi-path effects from the material making up the buildings around you, whereby signals bounce off them so the receiver actually thinks it is somewhere in the buildings around. Another factor that can affect this is the number and position of satellites that can be seen at the time of logging; trying the route on another day may give better results.

What images/satellite-photos are used as backdrop?

We use a mosaic of the world compiled by NASA from the Landsat-7 satellite. You can get a hold of them from onearth.jpl.nasa.gov. They are published as a WMS Server, so you can easily get images from that server yourself.

We also use Yahoo! Aerial Imagery.

What makes a road belong to a city?

This is often asked by beginners. There should be a closed way marking the extend of the city with a "place"- and a "name"- tag as well as a sindle node with a name and place to mark where to draw the city-name. In the actual map such a shape does not exist for many cities, thus only the distance to the node that marks the city can be used in these cases. For exceptional cases an is_in can be used.

what shall I do for roads that have multiple values for a tag?

The database can do this but most tools do not. So comma-separating the values is used. e.g. nat_ref="B500,B550" for a piece road that belongs to both, the B500 and the B550 at the same time.

Rendering Questions

I have just made some changes to the map. How do I get to see my changes?

There are currently two layers on the slippy map between which you can switch by clicking on the topright (+) icon.

1) Mapnik: The Mapnik layer data is updated once a week on Wednesday from planet.osm. However, the map is not automatically rendered. In order to update the Mapnik layer, view the area you want rendered and the request will be submitted automatically when you view an outdated tile. When you view the area you're interested in, and see outdated tiles, a request is issued to the tilesgenerator which updates the tiles. If you revisit the same area/zoom later you should see new tiles (might need a refreshing of the browser cache).
The data extraction for Mapnik starts at 00:00 (Midnight) Wednesday morning, changes will be viewable roughly 12 hours later. To be sure that your updates appear on the Mapnik layer in the following week, you should plan to get them completed by Wednesday 00:00 GMT (US 5-8pm Tuesday, Europe/Africa 0-4am Wednesday, Central Asia 4-8am Wednesday, Australia/East Asia, 8-10am Wednesday)
2) Osmarender: The second layer on the drop down list, called osmarender, is updated by the Tiles@home project. When you upload changes, some of the changes are recorded and the tiles you have changed should go into a queue. The tiles then update within a few hours; usually less than one hour. The amount of data within a tile will change the amount of time it takes. The more data the more time. Some changes are missed though and tiles therefore may not update for a long time. Manually requesting these tiles is possible through the slippymap at informationfreeway.org (only if you are at zoom level 12; check the properties on the map image and look at the URL -- 1361.png means zoom 12, tile coords 2047,1361), or by means of Ojw's tilebrowser.
If your data is still not appearing and you are pretty sure that the one or the other renderer has done its job, then you might have a tagging problem. Check that:
  • all your ways are tagged with something appropriate that will be rendered (eg highway=unclassified)
  • your tags should be in lower-case, HIGHWAY and Highway will not work.

What is the map scale for a particular zoom level of the map?

The following table shows the nominal scales for each zoom level. Data from http://labs.metacarta.com/osm/

So, for example the nearest equivalent to an OS Landranger map at 1:50,000 is zoom level 13 (nominally 1:54,000).

Note: figures are rounded to millions from levels 8 to 2.

Table of Levels/Scales
Zoom level Scale as representative fraction
18 1 : 1,693
17 1 : 3,385
16 1 : 6,771
15 1 : 14,000
14 1 : 27,000
13 1 : 54,000
12 1 : 108,000
11 1 : 217,000
10 1 : 433,000
9 1 : 867,000
8 1 : 2 million
7 1 : 3 million
6 1 : 7 million
5 1 : 14 million
4 1 : 28 million
3 1 : 55 million
2 1 : 111 million

What Geotagging do you use?

See Geotagging.

License questions

You find answers to legal questions on our Legal FAQ