From OpenStreetMap Wiki
Jump to: navigation, search

Comments about a new way to represent ruins have been move to Talk:Proposed_features/ruins

Historic Sites

How should I tag a feature (building or other POI) that has been torn down and there are no "ruins"? Perhaps something else was built on the same site to replace it. There are lots of imported nodes from USGS Geonames that have "(historic)" added to the end of the name, particularly school buildings that no longer exist. These nodes show up on the maps as if they did exist. I would rather not delete the nodes, but add a tag such as historic=site to mark the site of something that used to exist. Any opinions?--Elyk 00:19, 5 August 2009 (UTC)

I think we need some kind of life cycle tag to avoid the same problems with historic=ruins or ruins=yes. The other historic=* tags describe objects that are on the ground (buildings, shipwrecks, archaeological digs), although historic=battlefield and historic=archaeological_site don't quite belong unless they are preserved as historic sites.--Elyk 01:00, 5 August 2009 (UTC)

Reconstructed historic place

Here, in Ogden, Utah, there is a park called Fort Buenaventura. It is the location of the first, permanent settle of caucasians within the boundaries of current-day Utah. (It was established in 1846, one year before the Mormon settlers.) The original fort is gone, but they have reconstructed it as accurately as possible. I don't think that any of the "historic" tags applies. Is there a tag that would work for this? — Val42 01:14, 18 July 2009 (UTC)

I would tag it just the same as if it was the original. Maybe we should create an additional tag to mark reconstructions, e.g. reconstruction=yes or maybe including the year of the reconstruction. I am not aware of an existing tag for this. We have a similar situation here in Europe. There are many sites of roman origin, but not much has survived that long, almost all buildings are reconstructions. Also, some places marked as castles are actually ruins that were rebuilt in a modern romantic style in the 19th century. Currently these facts are ignored and the places are tagged as historic=*. So I guess your fort warrants a historic tag, too. And if we want to improve things, we can add a reconstruction tag. --Nop 04:44, 18 July 2009 (UTC)

Historic borders?

How about historic borders, like that of the Roman Empire, which as Limes is an UNESCO world heritage site spread across Europe? Or the German Empire until 1918, international and within Germany?--MattGPS 13:28, 12 June 2010 (UTC)

Maybe there's a need for another project like OpenHistoryMap? ;-) ----


What about historic towers that once were part of a city wall and with a city gate in between them?


historic=city gate?

Does it depend if the place today is known as "...gate" or as "towers of ..."? --OSMfan 20:34, 19 July 2010 (UTC)

I think a 'Historic=Tower' would be a good tag to make official as the OS StreetView layer has these features on. -- 08:44, 5 August 2010 (BST)


Is there any way to tag the period or age of a historic building? --Metaroll 15:06, 30 October 2010 (BST)

Nothing that is in common use (at least according to taginfo). Sometimes it's just the name that indicates the period. Perhaps historic:century=* would be the simplest way to solve this. --Mstriewe 21:25, 29 December 2010 (UTC)


I think we need to split country or city houses of a historic nature away from castle=*, since the usage documented there is a bit of a conflation in English. What about an ascending scale like:

  • historic=house Historically important ordinary houses, e.g. birthplaces of famous people.
  • historic=residence The smaller end of the scale for grand stately houses, e.g. Apsley House in London.
  • historic=mansion Very large stately residences, for dignitaries of less standing than those of a palace.
  • historic=palace Grand residences particularly used by heads of state or high-ranking dignitaries.

Just a suggestion. Any opinions? --achadwick 18:17, 13 April 2011 (BST)

Suggested introduction

Although it is barely included in the major renders, "historic" is one of the most widely used tags. It must be seen as having value by a large number of contributors. The values assigned mostly follow a similar pattern (noun for the value, categories named using a lower case & underscore character set, etc). The documented values are widely used, but there are also many undocumented values that follow a similar pattern. There are frequently inconsistencies in the way that things like changes of use, and current status are handled (ruins, abandoned structures, etc.).

As this feature seems to be of widespread interest it would be good if there was a way to encourage wider involvement, and more consistency in the way the tag is applied. It seems that too prescriptive an approach is not the best way of doing this. Only the most determined contributors have the stamina to wade through the current documentation and discussion on different options.

I would therefore like to suggest that this page should begin with quite a comprehensive set of general guidelines about the use of the "historic" tag, in a way that leaves scope for contributors to be expressive, rather than simply listing a very limited set of options.

Something like this.....

The tag “historic” is used to identify features that are of historic interest.

Each contributor is responsible for determining whether a feature or event is of sufficient interest to justify use of this tag. In selecting which historic features are significant contributors may wish to consider factors such as:

  • The degree to which the subject itself is considered to be of historical importance. Examples of supporting evidence could be inclusion in an authoritative listing, or a special protected status
  • The size of the feature, and in particular the degree to which it influences the visible landscape. The use of the “historic” tag with larger features, such as castles is relatively common, and the “historic” tag has also been used to explain the impact on the landscape of human activity, such as large-scale mineral extraction, or an abandoned canal
  • The benefits of highlighting destinations that are of potential interest to map users, such as historic sites that are now major tourist attractions
  • The degree to which the subject is of interest to the wider OSM community - the historical transport infrastructure, including roman roads, and disused railway tracks, is often identified in this way
  • The national and international significance of an important historical event, such as a major battle

The “historic” tag is always used to identify subjects that relate to human history. Geological features, and palaeontological sites (places containing the remains of ancient life-forms) are described using the “geological” tag. The "historic" tag should not be used for features that are considered an important part of the natural heritage, and should only be used or aspects of the cultural heritage that are also of historic importance. Where appropriate it can be used alongside other tags, such as "heritage", or "landuse"

The value in the historic tag is normally used to characterise the type of feature. Normal practice is to code this as a category. The range of different categories is open, but the preferred format is to use only lower case alphabetic characters.

Examples of the most common tags include:

  • historic=monument
  • historic=mine
  • historic=castle
  • historic=wreck
  • historic=battlefield

It is normal practice to separate words with underscores in the value of the historic tag. This approach is not universal in existing usage (for example the database contains many examples of both “citywalls” and “castle_walls”), but it is preferred because it improves readability. For example:

  • historic=mine_shaft
  • historic=roman_road
  • historic=wayside_cross

Wherever possible, contributors will wish to use a value that is already in common use. The most common, for different types of feature include:

  • Commemorative structures: memorial, monument, statue
  • Domestic structures: house, manor
  • Industrial structures: mine, mine_shaft, quarry, mine_adit
  • Maritime: wreck, ship
  • Military structures: castle, citywalls, fort, battlefield, castle_walls, earthworks, moat
  • Religious structures: wayside_cross, church, wayside_shrine
  • Pre-history: tumulus, stone_circle, menhir, standing_stone
  • Transport: roman_road, railway_station, bridge
  • Other: wall, boundary_stone, well, boundary_marker, folly

(for a more complete list of values that have been previously used, see:

A number of generic descriptions have been used where the specific type of feature is not known, or where it cannot be specified for some reason. Examples include:

  • historic=archaeological_site
  • historic=building
  • historic=industrial

Contributors should try to avoid common pitfalls such as:

  • Using the value of the “historic” tag to record either the operator of the facility (use the “operator” tag for this), or the name of the facility (use the “name” tag to record this).
  • The combination “historic=museum” implies that the museum building itself is the feature which is of historic interest. In some cases this combination has also been used to identify the location of collections of historic objects, but this approach is ambiguous. Use "amenity=museum" to identify important collections of historic objects, and "historic=museum" only where the building itself is of historic importance.
  • The combination "historic=monument" should only be used with larger structures that are built to remember and show respect to a person or to a group of people. The use of “monument” should be avoided for other types of structure, such as a church, and for smaller commemorative features the tag “historic=memorial” is preferred.
  • A number of historic features are no longer used for the purpose for which they are historically important. In this case the value applied to the “historic” tag records the original use, and should be used alongside other tags which record the current use. For example: “name=Durham Castle”, “building=university”, “historic=castle”.
  • The form “historic=yes implies that the type of structure is defined elsewhere. This is best avoided, because of the complexities involved in determining the type of structure reliably from other tags

--Peter Reed 19:06, 2 May 2011 (BST)

I see you've put an intro section on the page (Key:historic) now. That looks great. It's with just some of the above text picked out. As you've noticed, it doesn't make sense to list example tags and have the table displayed below. Your text above also goes into some more general tag principles which you've correctly missed out on the actual page. As I mentioned on your blog post, documenting general principles is something we've started doing linked off the Good practice page, so some of those ideas could maybe be worked in there -- Harry Wood 11:56, 5 May 2011 (BST)
Thanks Harry - I transferred the changes bit by bit, and tried to clean up the text as I went. So I'm not sure which version you commented on above. If I've moved into areas that I shouldn't have, please advise, and I'll be happy to fix (or, of course, welcome your changes). On duplicating the examples, I did wonder about this. My take was that there was some value in showing illustrative examples, without duplicating all the detail. I've tried to stick only to values that are very widely used, thinking that many contirubtors won't have the determination too chase down all the different sources of information, and the tables seem to be edited a lot - but hopefully the description should end up as being more stable. Learning as I go! --Peter Reed 12:35, 5 May 2011 (BST)
Ah yeah. I was commenting on an earlier version. Maybe it's looking a bit too long now (purely from a cosmetic point of view, taking a quick glance) But reducing the length of a wiki page is harder than writing text in the first place :-) -- Harry Wood 11:46, 6 May 2011 (BST)

Harry, thanks again for the advice. I have managed to trim it down a bit, and removed some duplication. Next steps get a bit more complicated.
I think it would help encourage consistency if it was easier for contributors to scan the the growing list of core values. These are not particularly organised at the moment, and it might be better to grouping these in a similar way to my list of common values (i.e. by Domestic structures, Military structures, etc.). This would involve a structural change to a template. I'm not sure I'm up to this, either technically, or in terms of understanding the courtesies / protocols. I'm happy to suggest it in the template talk page if this seems appropriate, but might have to rely on someone else to implement.
Similarly, the bit I have included in the intro on how to define new values might be better under the "User Defined" entry in the template. The same applies to the "good practice" suggestion on monument, and museum. Same issue with editing and protocols, but not quite as complex.
I think this page is still missing decent guidance on "ruins", "abandoned", "disused" etc. but this is more difficult, obviously of wider interest (I see you have analysed some of the options yourself), and seems likely to stir up something of a hornets nest. I get the impression that "historic=ruins" option is likely to be a good starting point, but I suspect I need to do a bit more background work first.

--Peter Reed 13:28, 6 May 2011 (BST)

Regarding the suggested values for pre-historic object there may be confusion with the tagging schema suggested at historic=archaeological_site. --Mstriewe 17:58, 8 May 2011 (BST)
Thanks Mstriewe, I have rephrased the section on generic values to cover this approach. --Peter Reed 22:39, 8 May 2011 (BST)

I disagree with "The form “historic=yes implies that the type of structure is defined elsewhere. This is best avoided, because of the complexities involved in determining the type of structure reliably from other tags." If a tree or something else that normally not is associated with history but nevertheless carries historical interest I think it is an excellent idea to add 'historic=yes' in the same manner as 'tourism=yes'. I think 'natural=tree + historic=yes' is more straight forward to render than 'historic=tree'. --RocketMan (talk) 17:02, 28 May 2014 (UTC)

Link to National Registries

I would like to tag items with their registration number at National Historic Registries. Any such tagging would need to identify both the registry and the registry's primary key for the resource. How about:

  • historic:nrhp=76000497

Which would be for Lyford's Stone Tower at

Brycenesbitt 02:22, 25 May 2011 (BST)

  • My feeling is that one might want to try and use key:ref in some way. In the historic:nrhp example, 'historic' and 'nrhp' are somewhat semantically redundant. Would in the example case there also be "historic=*" and "attribution=NRHP"? (Which I don't see in taginfo, but which I think would be better than the much used "US-NPS" in being more specific) One might then use "historic:ref=76000497". --Ceyockey 21:30, 26 November 2011 (UTC)
    Actually, looking at one of the NRHP records, the ref should probably be "ref:NRIS=76000497", as the code is attributed to the National Register Information System in NRHP records. --Ceyockey 21:33, 26 November 2011 (UTC)
    I've struck out my input above because it doesn't jive with the actual meaning of 'NRIS' ← it's a document designation for documentation supporting placement of a feature on the Register. Apologies for misunderstanding. Back to the topic at hand ... I've added attributes to a feature which illustrate one approach; see . In this case, the 'ud*' attributes were added by a previous editor; I added the following with the reasoning included below:
    A couple of alternatives include a) using 'historic:attribution' instead of 'attribution'; b) leaving out the NRIS reference altogether; c) using 'National Register of Historic Places' or 'US-NRHP' or 'National Register of Historic Places - United States' instead of the cryptic 'NRHP'. --Ceyockey 23:21, 26 November 2011 (UTC)

Formalising under a namespace, historic:, allows a generative scheme for different historic aspects of a single database object; see below. For example, a battlefield at which a later historically notable event took place (I don't much like Tag:historic=event conceptually, but it serves as an example). Since the refs above pertain to just the site and are current IDs/codes in some registry or database, I guess they should go under plain ref:<registry>. I've a suspicion that some refs might pertain only to one "historically notable aspect" but lie in the same registry and thus be susceptible to clash; if necessary, and if association is required, those could be bundled with that aspect as a historic[:<n>]:ref[:<registry>]=<regcode>, I suppose. This may well be overcooking it though, and IANAhistorian. --achadwick 12:37, 3 January 2012 (UTC)

Composite tag extension historic:url

see #Link to National Registries discussion above for context Proposing extension of the Core value set with the composite tag 'historic:url' with the meaning 'link to documentation supporting the designation of the historic item as of historic interest. Restrict to a single value.' --Ceyockey 23:34, 26 November 2011 (UTC)

Composite tag extension historic:attribution

see #Link to National Registries discussion above for context Proposing extension of the Core value set with the composite tag 'historic:attribution' with the meaning 'value indicates the agency, group or organization which has designated the historic item as of historic interest. Allow multiple values where multiple organizations have so designated the item.' --Ceyockey 23:34, 26 November 2011 (UTC)

Multiple historically interesting phases on a single feature

It may be the case that a single physical feature has multiple phases of historically interesting use, possibly with a current use too. Often (typically?) though, one historic aspect is more notable than the others. Since namespaces are being used for most extensions, would it be possible to represent that using a generative tag key scheme of the form

as current, but redefined to be the most notable historic aspect of a feature
generative form for tagging additional historic aspects

where <n> is a positive nonzero integer and <hvalue> is any value which makes sense for historic=*. Historically notable aspects can be kept sorted in descending order of notability for increasing <n>; note that the 0 slot is reserved so that it can be used in data consumers for the "most notable" aspect. This could also be generalised to subkeys of historic such as historic:era=* and its friends in a way that keeps the required association like this:

as current, but redefined to be the subkey associated with the most notable historic aspect of a feature
generative form for additional subkey/value pairs associated with historic aspect <n> for this feature

where <hsubkey> is any possible subkey for historic=* which is not an integer (such as the civilization string in historic:civilization=*), and <hsubvalue> is any value which makes sense for that subkey.

Does this seem useful and workable? It's admittedly a slightly complex schema, and this may not be the best way of explaining it, but it's nicely backwards-compatible with existing current historic use tagging at least. And of course it can be combined readily with current uses like tourism=* etc. without fear of misinterpretation: using the namespace will prevent that.

--achadwick 12:06, 3 January 2012 (UTC)

Sounds reasonable, especially in combination with subkey historic:civilization=*. However, this way of "namespace" tagging in general is very hard to understand for readers (both human readers and automated evaluation) in my opinion. Hopefully, we will get better data structures for that one day... --Mstriewe 11:35, 5 January 2012 (UTC)