Wikipedia talk:WikiProject Geographical coordinates/Archive 20

Default scale for {{coor dms}}?

In the case of coordinates including seconds, it seems to me that a scale of 1:300000 is too broad for such specific coordinates. Would it be possible/reasonable to change the default scale of {{coor dms}} to ~1:10000? Rami R 16:12, 16 August 2008 (UTC)

While 300000 might not be the best default. The issue is that coordinates and scale are independent variables and should be. One can give coordinates accurate down to the seconds - about 100 feet (30 m) for a large lake or reservoir, which is usually cited as the dam containing it or mouth draining it, so in that case, a small scale might not be appropriate. So if a scale is not specified, at least a "type" should be (which sets a default scale for that type). LeheckaG (talk) 20:14, 16 August 2008 (UTC)

Indeed scale and accuracy are supposedly independent, and cases with specific coordinates referencing large objects exist. However, I believe (I haven't actually tested this) that the in the majority of cases, when using specific coordinates one is referencing smallish landmarks. As such I believe that the default scale (when neither the scale or type parameters are given) should be significantly narrower than 1:300000. Rami R 21:38, 16 August 2008 (UTC)

Where to write the coordinates in an article ?

I added coordinates to Meidaimae_Station but I am sure I did it wrong, how am I supposed to do ? I could not find on the wikiproject page. Please tell me so that I do the next articles right. Thanks ! Nicolas1981 (talk) 07:07, 17 August 2008 (UTC)

User Wwoods gave me the answer :-) Just append "|display=title" in the {{coord ... }} tag and then the coordinates automagically appear in the upper right corner ! Cheers Nicolas1981 (talk) 09:05, 17 August 2008 (UTC)

To make life easier on yourself in the long-run, see Wikipedia:WikiProject Trains/Article templates#Article text and infoboxes. You probably should use either {{Infobox japan station}} or {{Infobox Station}}; or copy one of those or the Deutsch-German {{Infobox DB station}} or United Kingdom {{Infobox UK station}} template, and customize them as {{Infobox JA station}}.

A Coor(d) template can be placed in the Infobox in either a corresponding Place or Location fields. Coord ... "display=inline,title" ... causes coordinates to display both where the template is placed and the upper-right corner (article title bar). LeheckaG (talk)

Thanks a lot :-) I probably must use {{Infobox japan station}} as I am adding japan stations, and 50 existing one already use it. This infobox contains a lot of information but no coordinates, though... should I be bold enough and add a "coordinates" parameter like seen on {{Infobox DB station}} ? I am quite new in this area of Wikipedia, so I prefer to ask... is it legitimate to add the "coordinates" parameter to {{Infobox japan station}} ? Nicolas1981 (talk)

There are (3) possible approaches:

• Use an existing "Place" or "Location" field and put the coordinates are the beginning followed by <br> the-normal-WikiLinked-text-for-the-location

i.e. not sure how Japan is civilly subdivided, but elsewhere (like Canada/US): Address (if known), City, County, Province or State

• Go ahead and add coordinate(s) fields to the Infoxbox. Personally, I like to use {{Location map}} within an Infobox, which requires a separate latitude and longitude. In WikiText, it is easier to "build up", a.k.a. concatenate, rather than parse, a.k.a. split apart; consequently from separate latitude and longitude DMS fields any combination which is needed can be built, including generating a {{Coord}} template. Because Coord accepts additional coordinate (region:, type:, scale:, source:, ...) and template (display, format, name) parameters; there should be a collective coord_parameters field, and fields for each of the others.
• If an Infobox template is protected or you do not feel comfortable updating it, then post on the Talk page for the Infobox template asking about your proposed additions, and see what the response is.

I prefer to use {{Location map}} within Infoboxes because it allows you to create or use one graphic map (.SVG, .PNG, .JPG, ...) for the overall general area and use coordinates to overlay a marker on the map. So it requires less overall effort than generating individual maps for each article, and the Location map gives someone a general idea where a place is located.

LeheckaG (talk) 13:41, 18 August 2008 (UTC)

Thanks LeheckaG for your insight :-) As suggested by your third point, I opened the discussion on Template_talk:Infobox_japan_station. I am hesitating between {{Location map}} and {{Coord}}:

• location map makes it easy to have a cute small map in the infobox, very convenient.
• coord seems to be the standard for locating wikipedia things, which means more third-party parsers and applications will be programmed to understand it. Why use this other way of locating thing, is it legacy ?
• The best would probably be to make the small map builder understand the {{Coord}} format.

By the way, I encountered something that surprised me: how come some articles use both templates ? (To see one, edit Marbletown, New York and check "Pages transcluded"). Nicolas1981 (talk) 08:46, 20 August 2008 (UTC)

The Marbletown article you cited uses {{Infobox Settlement}}, which in turn transcludes {{Geobox coor}}, which in turn transcludes {{Coord}} which in turn transcludes a few "sub"-templates including {{Coord/link}}, which in turn transcludes {{Coor URL}}. This is somewhat "normal" behavior for Wiki templates, where transcluded "sub"-templates each do a piece of the work, and there is "reuse" a.k.a. borrowing/sharing functionality between template libraries/families rather than "re-inventing the wheel". In some cases, it could be re-done cleaner starting "with a blank slate" provided one knows and remembers the past issues and history involved in the development of the predecessors (so that one does not run into the same issues or make similar "mistakes"). In general, on Wikipedia, when something works - people generally re-use things as is and do not go tearing older things apart unless they "break" or the available functionality needs to be extended. Recently, I have gotten into the "habit" of verifying template source code against the documentation and trying to ensure that the template documentation is updated before using a template (at least that all parameters are documented along with their default values).

There are some articles which actually do use multiple Coor(d) templates: in the Infobox, in-line in the article text, and often in External links section. These separate coordinates often use different coordinates templates, and sometimes have inaccuracies where several different coordinates are stated for the same geographic feature. In general, coordinates should be "confined" to the appropriate fields in the Infobox(es) or in appropriate tables (when the article details several "related" features), and the article's ONE "primary" latitude,longitude coordinates should display in the article's title bar (in the upper-right corner).LeheckaG (talk) 13:42, 20 August 2008 (UTC)

LeheckaG, thanks a lot for taking the time to explain this in details ! I spent my day yesterday investigating, and will continue analyzing how pages are built, that's very interesting :-) I makes me reevaluate the strategies for writing a wiki geodata extractor. Cheers, Nicolas1981 (talk) 04:11, 21 August 2008 (UTC)

For geodata extraction, your two possibilities are either:

• ("internally", within Wiki) retrieving "raw" unrendered Wiki pages, and looking for the some of the variations of either coordinate templates or other templates like Geo-/Info- boxes, which transclude coordinate templates behind the scenes, and corresponding coordinates fields.
• ("externally", outside of Wiki) retrieving rendered Wiki pages and looking for Geo microformats. Which is what external search services like Google do.

On Wiki, like on ToolServer with PHP code scripts, you have both options available - there are some PHP scripts which can search for the rendered geo microformats, and some which can search the unrendered source pages for specific coordinate templates or coordinate fields.

Because searching text can be taxing on the Wiki database servers, bulk searches are best performed on an exported copy of the database and best performed off-line on another server. If coordinate searches can be more confined in scope, then they can be performed in the Wiki environment (usually on "ToolServer", which is actually more than 1 server, using recently replicated local copies of the Wiki database). LeheckaG (talk) 06:57, 21 August 2008 (UTC)

Here is DBpedia's GeoExtractor, which I was considering improving, it is already quite smart though. It uses what you called "internal" extraction, and additionally it looks for "coord" in each of the templates that would be transcluded during the page build :-) Nicolas1981 (talk) 14:00, 21 August 2008 (UTC)

Polar co-ordinates

See List of research stations in Antarctica, Extreme points of the Antarctic and Extreme points of the Arctic. For obvious reasons, distances are horribly distorted in the polar regions when viewing such areas. Is it possible to link to map services that do a polar projection? Are such services available? (Cross-post from Template talk:GeoGroupTemplate). Carcharoth (talk) 02:50, 18 August 2008 (UTC)

Thanks everyone for your work !

I am a huge DBpedia fan, and coordinates are precious for building impressive software based on Wikipedia infoboxes data... so kudos to everyone who makes the effort to contribute coordinates, it is immensely useful :-) Nicolas1981 (talk) 11:59, 18 August 2008 (UTC)

Geolocation search tool

I’ve create a search tool that allows for regular expression searching on the GeoHack links in the external links table. This has the advantages of near real time information and powerful pattern matching. The following are some example queries I have created as a demonstration of the flexibility of the system.

Description	MySQL Regular expression query
Sample searches
Coordinates imported from the CSWiki	_source:cswiki
Cities whose populations are under 1,000	_type:city_*\([0-9]{0,3}\)
Antarctic Circle (approximate)	=(66_[3-9]|66.[6-9]|6[7-9]|[7-9][0-9])[0-9_.]*_S_
Locations near Manhattan, NY	=40(_4[2-9]|.7|.8[0-5])[0-9_.]*_N_(73_5[5-9]|73.9|74_0?[0-2]|74.0)[0-9_.]*_W
Error detection
Bad template parameters	\{\{\{[3468]\}\}\}
Articles missing metadata	_[WE]_+($|&title|\{+[[:digit:]]\}+)
Missing W/E	_[NS]_[^WE]+_*($|&title|\{+[[:digit:]]\}+)
N/S mixed up with E/W	_[EW]_[0-9._]+_[NS]_+
missing ":" in parameters	_(region|source|scale|globe|type)_+
scale containing letters	scale[_:]+[^[:num:]]

— Dispenser 16:05, 19 August 2008 (UTC)

Can you make the wiki configurable. This would be useful to find heading:? locations on commons. --Dschwen 16:55, 19 August 2008 (UTC)

You can append &site=commons to the end of URL, but it isn't robust yet. Many assumption where made about the English Wikipedia as to speed up development. — Dispenser 18:10, 19 August 2008 (UTC)

Good tool. I used to fix a few erroneous coordinates. At some point, I suppose we should fix all the Greek coordinates, they fill up the result pages .. -- User:Docu

My bot fixed most of the Greek coordinates. I added your table to WP:GEO#Coordinates_search_tool with a column to note the last time they were fixed. -- User:Docu

I've been working a tool over the past few day that will parse all external links and dump them in database accessible to everyone on the Toolserver. But during the course of development I found coordinates that geohack would happily accept without throwing errors like python does, some of them that I haven't work in exception are listed in the log's output. And those greek coordinates unsubsting the templates I saw eailier? — Dispenser 21:48, 23 August 2008 (UTC)

So I guess its time for some interesting stats. First off, the most popular globe other than earth is the Moon (424), followed by Mars (70), then our sister planet Venus (14), then by Mercury(13). None of the gas planets are listed nor the moons beside ours. The most popular source for data label is GNS-enwiki (32139), dewiki (5868), gnis (3548), the UScensus1990 (2654), GNIS-enwiki (1432), and again by enwiki-GNS (1423). City (236803) is the most popular type followed by NULL (164795), and rather vague landmark (75239). Popular errors in the region type are adm2 (2665), admin2nd (904), city(348991) (206), country({{{area}}}) (55), city(300) (55), adm3rd (52). — Dispenser 22:40, 23 August 2008 (UTC)

I tried to find type:city\(348991\), but without success. I must have misunderstood your comment or it's already fixed since. -- User:Docu

My mistake, It should have been type: not "region type" and the parenthesis should be escaped. — Dispenser 18:07, 24 August 2008 (UTC)

Thanks. I forgot about escaping .. anyways, looks like the Brazilian coordinates are ok, except for the type and source part. -- User:Docu

type:city(348991) was changed to type:city. -- User:Docu

75239 landmarks .. -- User:Docu

Similar to wp-world/info.php, would it be possible to compile a list of all types and region codes that are being used? -- User:Docu

500 most used values for type and same for region. — Dispenser 18:07, 24 August 2008 (UTC)

Thanks. BTW in the meantime there are 1500 type:edu -- User:Docu

Just wondering, should I convert the following coordinates:

• 2665 "type:adm2", 904 "type:admin2nd" to "type:adm2nd"
• 602 "type:town" and 272 "type:village" to "type:city"

The later types are consistent with WP:GEO#type:T -- User:Docu

"type:admin2nd" seems to be used only in Thailand. It can be fixed next time an update is made there.

"type:adm2" seems to be used for US counties with source:UScensus1990 by AnomeBot. Possibly we should change all these to "type:adm1st". -- User:Docu

GEOnet Names Server

I'm trying to sort out just how accurate GEOnet Names Server is. I figure members of ths WikiProject should have some opinions, at least as far as coordinates are concerned. Currently, over at Wikipedia talk:Notability (Geographic locations) the conversation has turned into more of a WP:V debate, and I've suggested that using GNS and sources such as FallingRain that take their data from GNS is unwise. I've also opened an AfD on such an article; Wikipedia:Articles for deletion/Əngəlan. I'd appreciate any insight you folks might have, even if you reject my AfD nomination I'd feel better about it if it has wider consensus. Phlegm Rooster (talk) 06:00, 20 August 2008 (UTC)

In general, most government-run GNIS/GNS services (USGS GNIS, NGA GEOnet Names, CA BC, ...) are reliable and verifiable provided one understands the accuracy limitations of particular coordinates: degree = about 60 miles (97 km), minute = about 1 mile (1.6 km), second = about 100 feet (30 m). For a more accurate conversion, 1 minute = about 6,076.1 feet (1,852.0 m). In general (some minor clerical errors aside), government-published coordinates are accurate within +/- 2 units prior to any missing fields or zeros. The +/- 2 units (rather than +/- 1/2 unit) comes from (3) possible sources of "error": conversion from different coordinate datums, "rounding" or "truncation", and "deliberate" offset (for "security" reasons). The majority of coordinates are accurate within +/- 1/2 to 1 unit, the ones with the greater +/- 2 unit error margin should be in the minority.

For the (Azeri/Russian) language/(Arabic/Cyrillic) transliteration "issues" with the (Əngəlan, Angalan, Ангалан, Angilan, Ангилан, ...) village article please see my posts on the AfD discussion for the AfD-nominated article which you cited above.

On the other hand, for "Third-party" GIS/GNIS/GNS systems, what is needed is "Traceability", to know which "authoritative" or "official" source they received their information from, which some systems provide and others do not. I suppose "photogrammetric" (spelling?) verification is sufficient to verify coordinates provided by a third party (i.e. opening them up in a aerial/satellite map service and verifying that placemark on the imagery matches the described location), but for "encyclopedic" purposes, a better "authoritative", "official", or "primary" reference source for the coordinates should be cited when available. In general, I recommend MENTIONING third-party GIS/GNIS/GNIS systems for additional information, but I do not NOT recommend CITING them as the "official" reference source; i.e. it can often be the case that one needs to "go through" a third-party system to locate the "clues" necessary to find an "official" source. The third-party source should be mentioned as "informative" while the official source should be cited as authoritative/normative. Personally, I prefer separate "Notes" and "Reference" sections, and a "General references" sub-section under references. "Notes" are used to otherwise clarify information or text in the article, Geo/Info-boxes, or tables. For instance, to describe how (2) or more cited-reference facts "add up" to information in the article, or to otherwise document a "jump" between a cited reference and the article. "References" should be used only for cited i.e. <ref></ref> {{Reflist}} references. Whereas, "General references" can be used with {{Refbegin}}, {{Cite}}, and {{Refend}} to mention those sources generally used to write the article but not specifically cited, or to document useful references which could be used to expand the article in the future - and in doing so those General references would then be cited and moved from "General references" to "References". So a third-party GIS/GNIS/GNS system "reference" could be mentioned in either "Notes" or "General references", but an "authoritative/official" source should be "cited" under References. When possible, a "one-click link" to the exact reference page should be provided; for instance, use {{Cite gnis}} or {{GEOnet2}}, rather than {{GR}}.

For some places, there might not be any "coordinates" published by an "official" source, but rather a narrative description; for such locations, the "official" source for the narrative description should be <ref>{{Cite}}</ref> cited, and then a {{Cnote}} in a "Notes" section and {{Cref}} where the coordinates are used to describe the process employed to translate the description to the coordinates used in an article. For instance: "Google Earth (or Google Maps ...) used to geocode intersection of Route ... and ... Street, then ... manually plotted based on official description: ... from intersection. LeheckaG (talk) 16:50, 20 August 2008 (UTC)

Is it truly reliable([[GEOnet Names Server, I mean)? Looking at stuff I know about, it is obvious that GEOnet has the same serious errors fallingrain.com has (is there some discussion about blacklisting fallingrain, by the way?). "Populated places" are in some cases just names of streets, not hamlets or anything similar. Fallingrain then copies these data, and adds some fanciful population figure (apparently based on a radius of seven kilometres). Then people start arguing (like in the above mentioned AfD) that a place is verifiable, since it is mentioned in GEOnames (which is a dubious indicator) and fallingrain, which is an even worse copy of it. For Belgium, I wouldn't trust it one bit (just ask some people from Belgium to do some random check on the populated places provided by GEOnames, and ask them if many of those are actual villages or even hamlets, or just streets and so on). I don't know about other countries, I know that fallingrain is not reliable for the USA, but GEOnames does not include the US so that must have another source. I would prefer to not use either of them (or any of the many websites that use them as a source) as evidence for notability or even existence of a "named location" in the Wikipedia sense (a village, hamlet, town, community, ...). Fram (talk) 11:46, 25 August 2008 (UTC)

As to U.S. DoD GEOnet Names (NGA.Mil) database reliability - it is only reliable to the extent of information which they provide directly: "official" Anglicized version(s) of names, country and administrative region (province/state) containing the feature, and coordinates - which are only as accurate as they are willing to publish to the general public +/-2 in the last units.

For the U.S., the "standard" authoritative/official source is the corresponding USGS GNIS (GeoNames.USGS.Gov) database.

For Canada, there apparently is not a national system, but there are several provincial ones, like British Columbia has http://ilmbwww.gov.bc.ca/bcnames/.

I am not sure what Australia or the United Kingdom have? The U.S. DoD GEOnet Names database is supposed to coordinate with official sources in other countries, so I am not sure which official governmental sources they use. For named marine or nautical features it is a bit simpler because international law codifies the official authoritative source as being a country's Coast Guard or Navy equivalent.

I agree that third-party republishers should NOT be cited as an authoritative source, but only as a "informative" source - the difference between a "cited" reference and a general reference. The purpose of a "general" reference is so that people researching a topic can find "clues" as to more sources of information; versus a cited/footnoted reference is an authoritative/official/reliable source where specific data in an article came from.

Any source needs to be "taken with a "grain of salt", errors are possible, so Wikipedia contributors should not "blindly" copy "official" government published information without a "common sense" fact-check to verify that what is being included in Wikipedia "makes sense". For example, in Alaska the U.S. Census Bureau uses fairly large census geographic areas, and their cited coordinates for a "populated place" can often be "in the water" or "on a mountain" - so "common sense" is required to verify by actually comparing posted coordinates with an area map.

In geographic names databases, generally "street names" are not included among locales or populated places unless they are somehow "noteworthy". They do not necessarily need to be a civil subdivision (municipal corporation/government) but they do need to be a named "populated place" where the surrounding area is generally known by that name.

The names in GEOnet names are "Anglicized" so they often vary from the local name in the local language - so that needs to be taken into account.

There are different classes or types of NGA populated places:

• PPL - Populated Place
• PPLA - 1st-order administrative division seat
• PPLC - political capital
• PPLL - locality
• PPLQ - abandoned
• PPLR - religious
• PPLS - places
• PPLW - destroyed
• PPLX - section

so one needs to consider what kind of "Populated Place" one is referring to. LeheckaG (talk) 13:11, 25 August 2008 (UTC)

E.g. GEOnet gives the Belgian populated place (PPL) of "De Lieve Dochter", for which I can find no evidence at all. The same goes for "De legehaar" or many other Belgian "placenames". E.g. Wittingstraat is a PPL according to GEOnet, but is just a random street in Stekene[1].

There is no evidence that this street (or many of the other "straat" places in the database) are used to indicate the surrounding area. You can read Wikipedia:Articles for deletion/Polfbroekstraat for another street which can be found on GEOnet where the consensus was that it was just a minor street without any justification for an article. I would suggest to not trust GEOnet for any Belgian location, and to check some other countries (preferably with editors from those countries) to check if they used a particularly bad source for Belgium, or that the site is generally unreliable (even though it is government operated). Fram (talk) 13:50, 25 August 2008 (UTC)

The Belgium source for the U.S. DoD NGA GEOnet Names is http://www.ngi.be/; on Wikipedia: Belgian National Geographic Institute. In general NGA sources are http://earth-info.nga.mil/gns/html/relatedlinks.htm some of the links are out-of-date, but the source names are generally correct. When one is able to determine the foreign official source (like BE NGI), then when possible, that official foreign source should should be used instead of U.S. DoD NGA GEOnet Names.

In the WikiProject Geographical coordinates/Archive 20 at GEOnet Names Server or WikiProject Geographical coordinates/Archive 20 at GEOnet Names Server, Polfbroekstraat (Dutch: Polf "pants" Street) case, it is a street/neighborhood of Sint-Lievens-Houtem about 3,000 feet (910 m) West of the NGA published 50°55′00″N 03°52′00″E. NGA "Anglicizes" and republished foreign official sources, so apparently Belgian NGI "coded" Polfbroekstraat as a "Populated Place", which they (Belgian NGI) probably should have either clarified or coded differently. One would need to "go through the history books", it is likely/probable that "Polf broek" formerly was a small community or notable feature of some kind for which the current modern street is named.

In the WikiProject Geographical coordinates/Archive 20 at GEOnet Names Server or WikiProject Geographical coordinates/Archive 20 at GEOnet Names Server, Wittingstraat (Dutch: Witting Street) case, it is a street/neighborhood about 2,000 feet (610 m) South and West of the NGA published 51°12′0″N 4°0′0″E. In between cities/towns of Stekene, Klein-Sinaai, and Koewacht. Same BE NGI comments apply.

In the WikiProject Geographical coordinates/Archive 20 at GEOnet Names Server or WikiProject Geographical coordinates/Archive 20 at GEOnet Names Server, De Lieve Dochter (Dutch: "The Dear Daughter") case, is somewhere within 1 mile (1.6 km) of 50°54′00″N 03°29′00″E near Kruishoutem and Zulte.

Belgian placenames like "Polf Pants" street, "The Dear Daughter" ... are typically named after such and then the spaces are dropped to form a name.

So "common sense" is required, NGA confirms that such a "populated place" existed and publishes its coordinates to whatever accuracy (minutes are about 1 mile (1.6 km). Using the coordinates and looking at a on-line map service (within a few thousand feet or several hundred meters) one can see that it is likely a street/neighborhood name rather than a "town", "village", "city" ... so:

• It is necessary to use common sense.
• NGA only provides "evidence" that a "similar" (possibly miscategorized) feature with a similar name (keep in mind NGA names are "Anglicized") was or is "near" their published coordinates.
• NGA does not provide evidence of what exists there now.

So in the Belgian cases, it helps to restrict Google searches with "site:.Be", and also helps if you translate the names from the foreign language (Belgium names can be a variety - not just Dutch) to English, breaking it apart into the roots like "Polf Broek Straat" so that you have some more insight into what you are looking for. LeheckaG (talk) 16:43, 25 August 2008 (UTC)

Thanks. A "broek" is a swamp in toponyms, and should not be translated as "pants"[2]. Not that that changes anything. I agree with your analysis, but I just wonder why NGA is then considered a reliable source to indicate that a village ever existed (like at the AfD mentioned at the beginning of this section). Once we have other (independent) evidence that some place exists, we can use GEOnet to give us the coördinates and so on. But to create and/or keep articles solely on the strength of GEOnet seems dubious. Fram (talk) 18:58, 25 August 2008 (UTC)

NGA GEOnet Names can be used as a WP:RS to indicate the association between a Name and coordinates or where. As to the specifics of exactly what the place was or is, one needs to use and cite additional references. So that is where one needs to start researching history and other alternative sources. USGS GNIS is a little more reliable with regard to the what, since it is not relying on foreign official sources, so USGS GNIS (Populated Places, Locales, Historical, ...) can be more "standalone". As the Belgian examples demonstrate, foreign definitions of a populated place vary. LeheckaG (talk) 20:23, 27 August 2008 (UTC)

Reliable sources for coordinates?

There have been some discussion over at Talk:Camp 1391 of the Camp 1391 article should have coordinates, what coordinates, and if there are reliable sources for the coordinates and so on. Some help would be appreciated. We have a description saying it is "situated in Israel's North District near Pardes Hanna, less than an hour's drive from Tel Aviv and close to Highway 65 between Hadera and Afula.", but exact coordinates would be nice to have. // Liftarn (talk) 21:47, 21 August 2008 (UTC)

Does Israel have something equivalent to the US's Geographic Names Information System? That would be the obvious WP:RS. Otherwise, perhaps there's an impartial mapping service? —EncMstr (talk) 21:55, 21 August 2008 (UTC)

Since it's a black site and it doesn't appear on official maps and it's airbrushed out of aerial photos I don't think it's likley they publish the coordinates just like that. We have some sources[3][4][5][6][7] the question is how reliable they are. // Liftarn (talk) 11:20, 22 August 2008 (UTC)

UK rail station coordinates

User:The Anome/Railway station coordinates lists geocoordinates for around 1400 UK National Rail station articles that did not have coordinates as of 2008-08-23. The articles are listed by their current Wikipedia titles, after best-effort redirect following and disambiguation. See the page itself for source data: whilst I've devoted best efforts to this, the usual disclaimers apply. -- The Anome (talk) 00:45, 24 August 2008 (UTC)

Above ( #New types (WP:GEO#type:T) ), I suggested to create a new type for these. I haven't added it yet to WP:GEO#type:T, but I'd like to go ahead now. What do you think? -- User:Docu

Omitting scale: prefix?

The instructions in the scale:N section states that the Scale: prefix can be omitted. I can not understand how this can be correct. What would an example of this look like? Perhaps what is meant is that this parameter is optional? __meco (talk) 08:53, 24 August 2008 (UTC)

Yes, it should read "parameter" instead of "prefix". -- User:Docu

I changed it accordingly. -- User:Docu

Coordinate parameters (region:, type:, scale: source:) can be omitted, but it is good to understand what occurs when you do, and thus the "value" of including all the known parameters.

• region: is an ISO 3166 code which (can) determine which set of maps are available for some mapping services. It is generally region:XX-YY a 2 character ISO 3166-1 alpha 2 country code followed by 2 to 3 character ISO 3166-2 province, region, or state code. Omitting region: can cause some performance issues potentially resulting in a timeout error message, because ToolServer.Org then does a database lookup to try to resolve blank or empty regions.
• type: was initially intended as a "shortcut" to hint at the appropriate scale:, type: defines relatively reasonable default map scales for most types, scale: overrides the map scale hint provided by type:, it has "unofficially" become a "type" used for data-mining types of coordinates, omitting type: results in the default map scale (usually 1:300000)
• scale: sets the initial map scale when a map is opened and overrides the hint provided by type:, most map services allow zooming in or out from the initial scale, but it is generally good practice for the feature and its general surroundings to be initially visible so that a user does not need to zoom in or out
• source: provides a hint as to where the coordinates came from like source:gnis, the hints are used by "bots" doing data-mining to determine whether to copy or otherwise transform or translate coordinates. For instance, coordinates labeled as source:gnis generally came from one "good" source, like USGS, so they should not be modified nor replicated more than once by "bots". It is possible for coordinates without a stated "good" source to get copied or updated by bots.

So, ALWAYS try to provide a region:, you "should" provide type:; recommended that you provide scale: and source: if known. LeheckaG (talk) 10:27, 24 August 2008 (UTC)

As Meko pointed out, taken literally, the instructions stated that one could write "10000" instead of "scale:10000". This is fixed now. -- User:Docu

No applicable type

When mapping a metro station, the closest types I can find in the list at the type:T section are Landmark or Airport, hardly applicable any of them. Aren't there more types in existence or can't we create more types? __meco (talk) 09:14, 24 August 2008 (UTC)

landmark should be used if there is no other applicable type. The new "railwaystation" (see #UK_rail_station_coordinates might do as well. -- User:Docu

Perhaps when it has been introduced into the template code (or wherever these types are defined). If anyone uses this type currently the map will show with a scale of 1:300.000 which is inappropriate even for very, very large railway stations. __meco (talk) 12:32, 24 August 2008 (UTC)

I added it to WP:GEO#type:T. Support by geohack should follow, hopefully. -- User:Docu

I don't know who/what geohack is, but is this an educated assumption on your part? Or are you just hoping somebody will notice your edit? __meco (talk) 17:43, 24 August 2008 (UTC)

Geohack generates the mapsource page that opens when you click on coordinates. To have it updated, one needs to ask Magnus_Manske. Preferably only once we agree about the name and the scale. --User:Docu

Well, I used 1:50,000 for the several metro station articles to which I have now added map coordinates using the scale: parameter and omitting the type: parameter. __meco (talk) 18:10, 24 August 2008 (UTC)

You can still set a scale to override the default scale. As type:airport has a scale of 1:30,000, the default scale for type:railwaystation should probably be smaller. -- User:Docu

Is there a type:building? --NE2 19:00, 25 August 2008 (UTC)

Not on WP:GEO#type:T. The current type to use would be type:landmark. However, already 163 coordinates use type:building and 137 coordinates use type:buildings. -- User:Docu

More cleanup

A series of pages have links that appear to be substitute old geolinks templates. Special:Linksearch/www.mapquest.com finds a few (2044). If a standard coordinate template is available, these may simply be removed. -- User:Docu

Is it possible to run the query and omit the *_Talk: namespaces? i.e. probably only the :En: main article namespace and possibly the Template: namespace article pages need to be updated. LeheckaG (talk) 16:21, 24 August 2008 (UTC)

Probably not directly on the special page. The feature used to be on the interface, but didn't work. In AWB one can build lists from Special pages and use "Linksearch/www.mapquest.com" as input, then filter by namespace. -- User:Docu