OSM defines itself as a geographic database where tagging is free and unrestricted. One of the motto is "tag what you see on the ground". But many tags present in the database are not visible from the ground and not related to geodata like phone numbers, faxes, websites, operators/brands, opening hours, etc.
My question is to know if OSM is just an open database accepting everything, including yellow page data ?
If not, what kind of information should be accepted or rejected by the community ? Is that not the role of the OSM Foundation to fix the limits ? Do we have a document explaining what OSM is not, like Wikipedia ?

asked 15 Oct '10, 12:31

Pieren's gravatar image

accept rate: 15%

edited 29 Aug '13, 11:01

I had rephrased this as a proper question. Can you NOT revert such changes or rephrase this as an actual question? It helps in keyword searching, better indexing by search engines and general readability. Thank you.

(29 Aug '13, 11:10) MagicFab

Based of what we often read about what OSM is not, here are some points:

  • a slippy map : OpenStreetMap does not provide rendered maps as finality. The map on the main page is rendered by Mapnik and hosted on OSM servers but just as a 'showcase'.
  • a repository for 'unmodifiable' data. Importing external data with the condition that they should never be edited (e.g. administrative boundaries) is impossible in OSM. It's the essence of OSM to not freeze data. If you want data unmodifiable mixed with OSM, create your own mashup and keep your data outside the OSM database.
  • an airspace or aerial map : all possible routes between the tens of thousands airports around the world are not welcomed in OSM because they are not fixed routes (see (1)).
  • a map containing historical data which are no longer physically present (see (2))
  • contour lines for similar reasons as airspace mapping (oneway imports, not modifiable, disturbing normal editing, see (1))

(1) airspace mapping thread in OSM mailing list (2) historical data thread in OSM mailing list

permanent link
This answer is marked "community wiki".

answered 19 Oct '10, 23:47

Pieren's gravatar image

accept rate: 15%

edited 08 Jun '11, 18:12

A reasonable start, but I have to say it clearly once again: we don't map routes between airports because there are no fixed routes (at least in europe). The only thing where something like routes or roads exist is the final approach to the runway.

(20 Oct '10, 08:35) petschge

Aerial routes are not less "fixed" than sea routes. But many sea routes are present in OSM (and not only for ferries). So why what is accepted for the sea is not for the air ?

(14 Feb '11, 12:10) Pieren

it is accepted for sea routes because of 2 reasons: 1. The sea is empty in our renderings and editors. It doesn't distract mappers. Instead of drawing monsters we draw ferry routes to fill the voids. 2. Most maps do it. (Probably because of 1.)

(04 Mar '11, 00:47) dieterdreist

Adding to dieterdreist's answer, I'm quite sure that maps show sea routes also because they have been a very regular route for many many people for so many centuries. Sea routes also often connect roads, which air routes never do.

(27 Apr '11, 22:23) jaakkoh

also, sea routes tend not to connect every port with every other port, rather the ro/ro or passenger routes as marked on the map tend to be very static and just link to a single other port (e.g. Sconser/Raasay), or at most a handful (e.g. Portsmouth)

(13 Jun '11, 14:08) c2r

Answering just your final question "Do we have a document explaining what OSM is not?"...

As far as I'm aware the closest page we have to this already, is the "Good Practice" page which links to two or three other pages on specific points. I'd like to see that set of pages expanded upon. It's not specifically dealing with the "What OSM is not" kind of limits. I suppose we could call that the "Bad Practice" list, but I think the same page should accommodate both (rename it perhaps) Documenting in terms of things which are not included in the database could be helpful yes.

permanent link

answered 09 Jun '11, 11:32

Harry%20Wood's gravatar image

Harry Wood
accept rate: 14%


Wikipedia is has over 100 million readers and covers many controversial subjects. They adopted the policy to protect their brand. They do however have a close relationship with Wiktionary and the other Wikimedia projects. Some of those projects, like Commons, have very low standards as to what they accept.

So perhaps it's more appropriate to compare OSM with Wikimedia. We could place all the phone numbers into a separate database called 'OpenPhoneBook', or we could just use the existing infrastructure.

In some ways we are however more like Perl (or any programming language) :

  • Users contribute to the database. Perl programmers contribute to CPAN.
  • The tagging system is frequently updated. The Perl language is also being updated, especially if you include the libraries.
  • Changes to the tagging system may in some cases require coordination between the users and the application writers e.g. splitting the barriers out of the 'highway' key. All programming languages need change management as they evolve.
  • We have different applications. Perl has many interpreters.
  • We have State of the Map. They have State of the Onion.
  • We have face commercial pressures. So does any programming language.

Some people want to limit the scope of OSM to the data itself and exclude applications. The first problem is that there exists no approved standard for tagging anything and having reference implementations will fill the gap. The second problem is that many contributors wait for the arrival of applications that support certain tags before they contribute that data. Furthermore many contributors insist on open source applications.

permanent link

answered 01 Nov '10, 00:32

Nic%20Roets's gravatar image

Nic Roets
accept rate: 6%

edited 06 Dec '10, 05:33

The average mapper thinks of the map as containing the same type of information found on their navigation device, such as POIs. Taking the concept a bit further, they envision the additional information as useful for offline searches with a portable smartphone: Give me all the X within Y distance of my location. After expanding the concept, the query becomes "give me all the X within Y distance of my location which are open now". Because the data reside in the same database, results can be nearly instant, even when the smartphone cannot connect to the internet. As the search result pops up, the end user can click on the phone number to call, or click on the web site for further information.

Mappers who collect information often see phone numbers and even web sites posted on the storefront windows of some POIs. They then enter this into the OSM database - which is a single convenient point of entry. POIs and businesses could be placed in a separate database, then combined later into the smartphone device or renderer for ease of use. This would require developing the separate database, possibly requiring a separate step of integrating POI edits into all editors to recognize separate databases, with the only end result to reduce the size of the OSM database by a small amount.

OSM has already set a precedent of including imaginary data with the inclusion of administrative boundaries. While I don't know of a list of what OSM is not, the verifiability page is a step toward filtering out completely unrelated information.

permanent link

answered 15 Oct '10, 15:21

Mike%20N's gravatar image

Mike N
accept rate: 17%

So if the list of ingredients of my preferred pizza from my well known pizzeria is "verifiable" (listed in the menu), I can enter it in OSM ? If the fuel oil price is "verifiable" (large display in front of my gas station), I can enter it in OSM ?

(15 Oct '10, 15:55) Pieren

The fuel price is "verifiable" but changes so fast that it's not worth recording in OpenStreetMap. The relevant time scale here is the time between planet dump (once a week) plus the time for a planet import (another week). So anything that changes faster then once every couple of weeks is definitely NOT worth having in OpenStreetMap (unless it's periodic).

(15 Oct '10, 21:19) petschge

You're right that if restaurants published their menus as XML exports, that someone would want to upload the XML menus to the map. Currently, the thing that stops people is that the amount of effort to collect and enter menu information versus the possible benefit: very little; if the web site is listed, people should just go there for more information.

(16 Oct '10, 12:26) Mike N

The philosophical ones can also be fun.

Mike N has just given some great and practical example of what the "extra" data can do.

Now if you are thinking towards "fixed limits" you would either explain next why missing limits are somethings negative or dangerous. Or you could give practical examples for the benefits of such limits.

Maybe you are concerned about server capacity or finances? In that case, I would suggest a simple appeal to those who love to tag lots: "If you enter lots of data then please donate lots of money for more servers."

I love the fact that such a young system can still function without many rules. In the long run, where many people interact there arises a need for regulation. But maybe in a context like a lovely world-map modern technology can allow for much of our personality quirks and our temperaments.

Just let us know more please, what your concerns are. Or maybe you just enjoy rules? I guess it is beneficial for a community to have some of each flavour. :)

And as a beginner, I actually like your question about a document. This is a precise question that one of the veterans should please answer with a yes and a link or just a no.

permanent link

answered 16 Oct '10, 03:16

Screentoosmall's gravatar image

accept rate: 0%

edited 16 Oct '10, 03:18

Your answer
toggle preview

Follow this question

By Email:

Once you sign in you will be able to subscribe for any updates here



Answers and Comments

Markdown Basics

  • *italic* or _italic_
  • **bold** or __bold__
  • link:[text](http://url.com/ "title")
  • image?![alt text](/path/img.jpg "title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported

Question tags:


question asked: 15 Oct '10, 12:31

question was seen: 13,030 times

last updated: 29 Aug '13, 11:10

powered by OSQA