When discussion has ended, remove this tag and it will be removed from the list. If this page is on additional lists, they will be noted below.
Maps are used as references in 32,000+ articles. From time to time questions about their use are raised in venues such as WP:GAC,WP:AFD and WT:OR. Policy and guidelines about sourcing and verifiability do not directly address nontextual sources. This RFC was started to answer some of those questions. I feel the Wikipedia community would benefit if we have some codified guidelines about their use to avoid having to continually revisit these topics.Dave (talk) 05:29, 19 March 2023 (UTC)[reply]
Source information does not need to be in text form—any form of information, such as maps, charts, graphs, and tables may be used to provide source information. Routine interpretation of such media is not original research provided that there is consensus among editors that the techniques used are correctly applied and a meaningful reflection of the sources.
Support—we allow foreign-language sources, and this would make explicit that non-textual sources can be translated into prose for articles. Maps are regularly cited as source material in other works, so they should be good for Wikipedia purposes too. Imzadi 1979→06:20, 19 March 2023 (UTC)[reply]
Support - The purpose of the original research policy is to prevent original ideas that have not been peer-reviewed or subjected to an editorial process from being added to Wikipedia, a reliable collection of facts. It is surprising to see the policy applied this way to the basic activity of reading a map, an activity billions of humans do every day and that is taught in schools worldwide. Quite frankly, it is shocking to apply the OR policy this way. Maps have been acceptable sources on Wikipedia for years and are used on several FAs (including recently on Coventry ring road). They are used on over 32,000 articles [1] across many different geographic subject areas, and Google Maps is cited thousands of times [2]. In fact, a similar statement in the WP:NOTOR essay was added in 2011[3] and nobody bothered to remove it from the essay.
The types of maps that are used as sources on Wikipedia are not drawn up from the recollection of one's mind - if it were so, like the maps that I drew as a kid, then that would be OR. They are not even the cartoonish ones drawn up by the local tourism bureau that would fail SPS guidelines. The ones we are talking about are from reliable publishers like Rand McNally, Michelin, Collins, and others that have been around for generations. Google Maps has been called out for supposedly being less reliable than traditional maps and for being unarchivable, and while a separate discussion can be had on that matter, I see no reason why they are not less reliable than other sources that we use on Wikipedia on a daily basis. Modern maps are made through taking GIS data (the primary source) and choosing what elements of that data to incorporate into the final product, upon which editorial processes are followed. In the days before GIS data, satellite photos were used. Map publishers follow an editorial process, just like all other sources we use on a regular basis (and resulting in a secondary source).
It seems that subconsciously we know that published maps are reliable sources. For millennia our ancestors have relied on paper maps for their daily livelihood. Even the online directional services that we use today are based on maps. Why do public libraries spend so much money archiving old maps if they are not reliable? We trust our lives to them, and yet somehow they are not reliable enough for Wikipedia. Such an attitude is intellectually elitist. Nevertheless, there is this idea that one needs special training to understand a map properly. I suspect there is a huge generational gap here. I can say that I have been reading maps since I was five years old, if not before. This is a skill that was in my elementary, middle, and high school curriculum in the years afterward.
Yes, there have been problems where people have read things into the map that are not there, and that is original research. And separate questions are below to examine the parameters of that.
Disallowing the use of maps as sources (which a no consensus result here would effectively do) will result in implications across several subject areas - really any article that has to do with geography. In the subject area where I edit, highways, it means that we basically have to remove 90% of the text that talks about the road as it is today - even for articles where there are literally hundreds of newspaper sources that bring the subject well past any reasonable standard of WP:GNG. Recently, there have been efforts to remove text cited to maps in highway articles, considering it to be "original research". I have to wonder why suddenly this has become such a big issue worth starting revert wars over: [4][5][6][7] This would be equivalent to having to remove entire plot summaries of movies and TV shows because there are no reliable sources describing them, as might be the case before sites like AV Club came about. (Which, by the way, are less objective than reading a map, since the symbolism in fiction is often quite subjective - frequently, I have to read Wikipedia just to understand what happened in a movie). It brings into question why we still allow routine calculations since the exact same language is in the above proposal. It means that by the same logic charts, graphs, and tables should also be banned, though apparently some believe that should be done too. How do these moves benefit the reader?
In short, declining this opportunity to affirm the use of maps and all non-textual sources would be a serious mistake for Wikipedia and I fear what such a move would say about the health and future of the site. Rschen775407:41, 19 March 2023 (UTC)[reply]
FTR, that addition in 2011 was by a perennial Wikilawyer - indeffed a couple of years later - who was trying to improve his position in a content dispute. A similar attempt to change WP:NOR failed. Notably, for example, here. Kahastoktalk22:27, 19 March 2023 (UTC)[reply]
I am (way too) familiar with Martinvl, but that does not mean that he was wrong on that occasion. And nobody bothered to remove it. Rschen775422:52, 19 March 2023 (UTC)[reply]
Oppose as currently worded. Sometimes, use of a map as a source will not violate WP:OR, but that is not always the case. I am concerned that if passed the current wording would be used to support the use of maps even when their use would be a violation. BilledMammal (talk) 11:22, 19 March 2023 (UTC)[reply]
But you agree that a map can be cited as a source of information in an article? If so, then you should support the inclusion of text that states Routine interpretation of such media is not original research provided that there is consensus among editors that the techniques used are correctly applied and a meaningful reflection of the sources. If something is misused in an OR-type situation, that interpretation is no longer routine nor supported by consensus. Imzadi 1979→15:48, 19 March 2023 (UTC)[reply]
See my comment below; I would support this proposal if it made it clear that this did not permit the interpretation of primary sources; is that a modification you would support? BilledMammal (talk) 16:06, 19 March 2023 (UTC)[reply]
I wonder if people have a shared understanding of what it means to "interpret" a map. I think that, generally speaking, we actually do want editors to confine themselves to "straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge" (e.g., "California is on the western side of North America", with the citation being to a political map of North America). I would not consider that to be "interpretation", but others might use that word for the act of reading a map. WhatamIdoing (talk) 04:30, 20 March 2023 (UTC)[reply]
Support – I think it’s a good move to copy the wording from routine calculations, since it’s a similar form of information translation, and relying on consensus to decide what’s "routine" and what’s not is the best we can do. Some editors may try to use this wording to support their WP:OR edits, but that’s true of all policies; as long as other editors are around to weigh in reasonably and in good faith it should be fine. Justin Kunimune (talk) 12:31, 19 March 2023 (UTC)[reply]
Support. The MLA Handbook for writers of Research Papers (7th ed., 2009) is specifically addressed to high school and undergraduate university students, who face more or less the same requirement to avoid original research that Wikipedia editors do. This handbook specifically recognizes the use of atlases, including Google Earth (p. 14). Jc3s5h (talk) 14:07, 19 March 2023 (UTC)[reply]
This is news to me, and interesting. Can you comment further on what the MLA now says? I have a few MLA handbooks from my college years, but I went to college before Google Earth existed. Dave (talk) 20:01, 19 March 2023 (UTC)[reply]
Looking at the MLA website it appears the publication I referred to has been replaced by MLA Guide to Undergraduate Research in Literature (2023) which I do not posses. The passage on page 14 I referred to states
Atlases are collections of maps. Along with the many useful atlases published as print volumes, prominent atlases available on the Web include The National Atlas of the United States of America, the official atlas of the United States; Google Earth, which covers the entire globe; and Perry-Castañeada Library Map Collection, at the University of Texas, Austin, a historical collection.
Support - Reading a map is the same as reading a foreign language source. You have to "interpret" symbols (glyphs) and reach a conclusion. Anybody can do the same and come to the same conclusion, otherwise the same principals apply that would to an English language text. - Floydianτ¢14:54, 19 March 2023 (UTC)[reply]
Support The edits will still be subject to our usual policies. If anyone sees a map and they want to contest its inclusion for some good reason, they will still be able to do that. Selfstudier (talk) 15:35, 19 March 2023 (UTC)[reply]
support Of course it depends on the kind of map, but essentially maps are just another way of presenting data. EDIT: Per Billedmammal below I want to clarify that my support is conditional on it being made clear in the text that WP:PRIMARY, WP:OR, and WP:V all still apply to maps. --Licks-rocks (talk) 15:43, 19 March 2023 (UTC)[reply]
Comment I could support this if Routine interpretation of such media is not original research provided that the source is not primary and there is consensus among editors that the techniques used are correctly applied and a meaningful reflection of the sources if the bolded text was added; I think we need to make it clear that this does not create an exception to our current policy on forbidding interpretation of primary sources. BilledMammal (talk) 15:51, 19 March 2023 (UTC)[reply]
I believe "interpretation" in the proposed addition has a different meaning than in WP:PRIMARY. WP:PRIMARY, in point 3, amplifies the meaning of "interpretation" as "a primary source may be used on Wikipedia only to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge." I believe the kind of interpretation intended by the proposed addition is routine map reading, and fits into the kind of statement that can be verified by any education with access to the map. Jc3s5h (talk) 20:02, 19 March 2023 (UTC)[reply]
Oppose as written, specifically provided that there is consensus among editors that the techniques used are correctly applied and a meaningful reflection of the sources. This will end up with projects or even article talk pages making up their own agreements. Maps can be used for referencing, but editors interpretations of details that do not exist on the map itself is still WP:OR. -- LCU ActivelyDisinterested∆transmissions∆ °co-ords° 16:05, 19 March 2023 (UTC)[reply]
This is the issue for me, it's the difference between something that is obvious from the map and something that requires interpretation. The moment that line is crossed it's not about whether it's a reliable source it has become OR. -- LCU ActivelyDisinterested∆transmissions∆ °co-ords° 18:44, 19 March 2023 (UTC)[reply]
However, by policy we have to interpret prose sources too. If we just copy them verbatim that's plagiarism. By policy, we have to summarize. Summarize requires interpretation. So the issue then becomes is can maps be interpreted correctly (assuming nobody here disputes that all sources can be interpreted incorrectly). I can think of several ways to argue that yes, visual source such as maps can be reliably interpreted correctly. Arguably the construction industry is tasked with interpreting maps (if we consider blueprints and engineering drawings a form of map) and they've been doing it for hundreds, if not thousands, of years. The fact that this world is full of structurally sound structures that look just like the blueprints, says that yes, it is possible to reliably interpret visual sources. Dave (talk) 19:22, 19 March 2023 (UTC)[reply]
I would argue that's exactly what this is about, summarizing sources. If sources aren't summarized correctly, do we blame the source or the summarizer? Dave (talk) 20:04, 19 March 2023 (UTC)[reply]
My point is that you can state what Shakespeare says in a sonnet and reference the sonnet, you can't state what he means by the sonnet and reference the sonnet. That would be your interpretation and would require a proper source. It is the latter that has been happening with maps, that has nothing to do with summarisation. -- LCU ActivelyDisinterested∆transmissions∆ °co-ords° 20:13, 19 March 2023 (UTC)[reply]
But if I look at a map of North America, and conclude the California is on the western edge, is that "just my interpretation"? I hope that you will agree that I have used the "correct techniques" (looking at the compass marking on the map) and that it is a "meaningful reflection of the source". WhatamIdoing (talk) 04:34, 20 March 2023 (UTC)[reply]
Oppose, this is the part that gets me "provided that there is consensus among editors that the techniques used are correctly applied and a meaningful reflection of the sources." because this will only exacerbate the problem we have of wikiprojects (like roads and highways, note the overwhelming brigading of this discussion by members of that project) making their own rules and pretending like it has community endorsement. I would also note as others have that doing our own interpretation of a primary source is strictly forbidden and this language as it stands would appear to provide room for those rogue wikiprojects to challenge the existing community consensus over primary source use. Horse Eye's Back (talk) 16:27, 19 March 2023 (UTC)[reply]
Sorry, I had assumed if you were opening this discussion you'd been following the roads and highways shenanigans. Isn't this an attempt to address those rogue standards around maps usage? A less charitable interpretation would be that this is an attempt to codify those rogue standards. Horse Eye's Back (talk) 17:09, 19 March 2023 (UTC)[reply]
Please remember that assume good faith is a fundamental principle on Wikipedia. Please stick to arguments about the subject matter, not arguments about your unchecked assumptions about peoples' motives. VC 17:28, 19 March 2023 (UTC)[reply]
I was asked a direct question, I responded. I have stuck to the subject matter. You on the other hand have not made a single comment addressing the subject matter, just this off topic one about me. Horse Eye's Back (talk) 17:30, 19 March 2023 (UTC)[reply]
Again, I would like to see arguments about the subject matter, not arguments about how someone else's statements are not relevant because they do not have a history of advancing arguments in an RfC less than 12 hours old. VC 17:45, 19 March 2023 (UTC)[reply]
I have contributed to the project pages of maybe 3 or 4 wikiprojects that use maps. I have yet to see any such wikiproject be sanctioned by arbcom (or whatever) for rogue behavior. If you are aware of wikproject that has been sanctioned for rogue activity please advise ASAP as this is something I would want to know about. Dave (talk) 17:34, 19 March 2023 (UTC)[reply]
I never mentioned sanctions. I don't think anyone disputes that the way highway and road route descriptions are sources is rogue, no other part of the encyclopedia sources like that. If they did there would be not reason to have this conversation. Horse Eye's Back (talk) 17:39, 19 March 2023 (UTC)[reply]
I appreciate that you retracted the words "rogue wikiprojects". However, IMHO deleting words from your post and pretending you never said them is poor form after people have already called those words out. It's OK for correcting typos or grammar errors, but IMHO not an appropriate way to retract. The more accepted way is to leave them in tact, but use strikethrough so people can see that yes, you did state those words, but now retract them or regret saying them. Cheers. My apologies. That was an irrelevent comment caused by confusing two threads open at the same time. Again apologies. Dave (talk) 18:56, 19 March 2023 (UTC)[reply]
Support Reading a map isn't OR for what is on it. If you can't understand the symbols there is even a legend. A map has the same pro/cons as any other RS and should be judged the same. Sammy D III (talk) 17:42, 19 March 2023 (UTC)[reply]
Support. Reading a map is not original research, just like looking at entries in a database is not original research. Maps collect data from a variety of sources and present that data in a visual manner. As others have said above, there are legends and scales which can help people . However, policies such as WP:PRIMARY and WP:V should still apply to maps. Epicgenius (talk) 18:24, 19 March 2023 (UTC)[reply]
Support: maps, tables, graphs, photographs, audio, video, and other non-prose elements are always in use as reliable sources. They are often primary sources, but is using that information more of OR than using information taken from written text? I don't see that at all. ɱ(talk)19:16, 19 March 2023 (UTC)[reply]
Support per Floydian and others. I don't see the need for a carve-out of primary source maps; WP:PRIMARY legitimates "straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge" and a great deal of map "interpretation" is unambiguous in this sense. Choess (talk) 19:48, 19 March 2023 (UTC)[reply]
Strongly oppose as the proposal is extremely misleading. The proposer claims that the text is adapted from WP:CALC, but that policy page does not mention charts, graphs or maps at all. This proposal is to "move" text, but in fact would add completely new text. While I don't necessarily disagree with the idea, I strongly object to the way it's been presented and feel that this completely disqualifies the proposal. pburka (talk) 21:19, 19 March 2023 (UTC)[reply]
You said you don't necessarily disagree with the idea, so there is value in continuing the discussion and not just invalidating the opinions of everyone who has already posted for something that is eminently fixable. Can you explain how you would correct the proposal to fix the problem with how you feel it is presented? VC 01:15, 20 March 2023 (UTC)[reply]
I think any changes to policy are premature at this time. There's clearly debate over the proper use of maps, but I'm unconvinced that WP:NOR is the right place to resolve it. Instead, I recommend someone start by drafting an essay explaining when maps can be considered reliable sources, how to determine if a map is a primary, secondary or tertiary source, and what information can be reasonably derived from a map without original research. The essay should quote existing policies, as I think they're already sufficient, even if they don't explicitly discuss maps. It should avoid carving out special exceptions for maps, even if there might already be local consensus for these carve-outs in some projects. pburka (talk) 03:26, 20 March 2023 (UTC)[reply]
Support If the map is a reliable source, then it can be cited in Wikipedia. The question of whether a specific map is a reliable source for specific information in Wikipedia is no different to the use of a book or website as a source and relates to issues like "who created/published it" and why we believe they are to be trusted to provide accurate information about that topic. Kerry (talk) 21:57, 19 March 2023 (UTC)[reply]
Only if it is genuinely routine, subject to a much stricter interpretation of "routine" than some seem to be applying here. If you're using the map to give spot heights or names for mountains, that might be genuinely routine. Interpretation of a map, such as giving a textual description of an area based on a general impression gained from contour lines and the like, is OR. I'd add that measuring land areas or distances between two points, even given a scale, is not a routine calculation and is thus OR (because the user may not adequately account for the projection). Kahastoktalk22:27, 19 March 2023 (UTC)[reply]
I consider routine map reading to be using the map the way the publisher or author intended it to be read. Some interactive maps have functions provided to measure lengths and areas. Some paper maps such as a sectional chart have a printed graphical scale on them. If it's well known that chart users, such as pilots for sectional carts, are expected to take distances off the chart, then taking distances off the chart is routine. Jc3s5h (talk) 23:06, 19 March 2023 (UTC)[reply]
CALC forbids calculations that are predicated on "specialized knowledge", which would include the distance information used by pilots. So that is most certainly not routine. JoelleJay (talk) 23:18, 19 March 2023 (UTC)[reply]
Actually, WP:CALC specifically says, "Mathematical literacy may be necessary to follow a "routine" calculation". Is mathematical literacy specialized knowledge? Dave (talk) 00:47, 20 March 2023 (UTC)[reply]
WP:CALC does not contain the phrase "specialized knowledge"; that phrase occurs in WP:PRIMARYSOURCE. That would seem to say, if read literally, that specialized knowledge could be used to read a map that's considered a secondary source but not a map that's considered a primary source. But I would not consider using a map the way the publisher intended it to be used as involving specialized knowledge. Jc3s5h (talk) 00:56, 20 March 2023 (UTC)[reply]
Creating a detailed flight plan from a sectional would not be routine map reading. Finding the distance between two airports that are on the same chart would be routine map reading. I watch 12-year-olds do it several times per year. Jc3s5h (talk) 01:31, 20 March 2023 (UTC)[reply]
You said "using a map the way the publisher intended it" doesn't involve specialized knowledge, which is totally untrue. And if the chart requires one to understand how to find distance data among numerous other symbols it is not routine. The average American adult cannot decipher a section map, and that is the threshold SKYISBLUE operates under. JoelleJay (talk) 05:44, 20 March 2023 (UTC)[reply]
(EC)Oppose. Maps are often primary sources. Unlike published paper atlases or road maps where the scale of a map and which details to include have been chosen deliberately and edited for accuracy by human authors, online database-generated maps with continuous spatial resolution have not received such oversight or secondary analysis at every zoom level. The latter group includes the GIS data-rendered dynamic maps from Google et al that are favored by highway projects and used to justify inclusion of arbitrary minutiae such as exact mile markers for individual localities and intersections along a route. Google Maps does not distinguish between cities and tiny non-notable communities (and so the choice for which ones to include on wiki will necessarily involve OR and likely circular referencing, with editors clicking on a Maps locale to bring up the linked wiki article and judging from its contents whether it should be mentioned), and these locations are not visible at every zoom depth; nor does it contain exact distance measurements between points (editors use various extensions for these calculations). For these reasons (not to mention known accuracy issues) the data extracted from such maps are not easily verifiable in the ways excepted by CALC nor are they DUE as they have not received secondary analysis. JoelleJay (talk) 23:15, 19 March 2023 (UTC)[reply]
Note that a significant chunk of those are citations to OpenStreetMap (such as at North–South MRT line) which is WP:USERGENERATED and in the process of being removed. In fact its the sheer number of citations which has made the process take so long. Existing citations do not support a consensus of reliability, period. For example Findagrave.com has thousands of current uses and we have a clear consensus that it isn't reliable. Horse Eye's Back (talk) 00:34, 20 March 2023 (UTC)[reply]
I don't think you know what consensus on Wikipedia means... in particular, the qualities of "not a vote" and "it can change". JoelleJay (talk) 01:06, 20 March 2023 (UTC)[reply]
Scott's correct, though: Most of our policies and and guidelines come out of actual practice. If a written page says "Don't do this" and experienced editors have done this in tens of thousands of articles, it's usually the written rule that needs updating (e.g., to "Don't do this unless..."). Wikipedia:Policies and guidelines begins with the words Wikipedia's policies and guidelines are developed by the community to describe best practices.... It is not an accident that it says they describe, rather than imposing rules from top-down. WhatamIdoing (talk) 04:45, 20 March 2023 (UTC)[reply]
I don't think you read what I or HEB or even Scott said. Just because some source has been used thousands of times in the past without the wider community noticing doesn't mean it's within policy, and it certainly doesn't mean a global consensus to eliminate it can't be developed. If "actual practice" was the end-all determinant our policies would've been changed to permit Carlos Suarez's water pump village stubs and the Daily Mail would be an accepted source. JoelleJay (talk) 05:54, 20 March 2023 (UTC)[reply]
Yeah, less than 6 months ago. Road FAs have been running on the Main Page for over 15 years. For a few years, they were doing them once a month until we told them to stop doing that. --Rschen775406:54, 20 March 2023 (UTC)[reply]
I can confirm that WhatamIdoing did indeed read what I said, because she made precisely the same point I was making (with more links and quotes, which are appreciated). Policy comes second after what the community is actually doing in practice, and it has been that way for at least the last eighteen years I've been here. —Scott5114↗[EXACT CHANGE ONLY]10:47, 20 March 2023 (UTC)[reply]
Just because some source has been used thousands of times in the past without the wider community noticing doesn't mean it's within policy – actually, it does. If hundreds of experienced editors do something tens of thousands of times, without the other thousands of experienced editors objecting, then that pretty much means that this is the community's policy in actual practice. Policy isn't written words; it's what people frequently do or recommend in a given situation. (Consider: "Honesty is the best policy", which does not mean that the word honesty was written down in a particularly high-quality fashion.) When we have hundreds of editors citing maps, in tens of thousands of articles, then we actually do have a policy of permitting citations to maps.
You are absolutely correct that Wikipedia:Consensus can change. I don't think this one has, or that it will, but if you want to achieve that, then now's probably the best opportunity to form a consensus against the established practice that we're likely to see during the next few years. WhatamIdoing (talk) 15:28, 20 March 2023 (UTC)[reply]
Not a single person has opposed permitting citations to maps. You are currently commenting in a section about the wording of WP:OR, not about whether or not we are permitted to use maps as sources (we obviously are). Horse Eye's Back (talk) 15:34, 20 March 2023 (UTC)[reply]
Let's go back to your belief that Existing citations do not support a consensus of reliability. Do you think that the individual editors who added those existing citations, and the individual editors who declined to revert those citations, actually believed that these existing citations were unreliable for the specific claim being supported? WhatamIdoing (talk) 15:40, 20 March 2023 (UTC)[reply]
In general on wikipedia we WP:AGF. Existing citations do not support a consensus of reliability, there is no policy or guideline which supports that and if someone tried to make that argument at RSN the community would laugh in their face. Horse Eye's Back (talk) 15:44, 20 March 2023 (UTC)[reply]
If something is used extensively by a few editors (or even by many editors), but later on it is determined by consensus that such use is not within policy, then that use is not within policy. Scott is arguing that being used in 30k articles means something is automatically acceptable; HEB and I are pointing out that that argument is incorrect using the example of findagrave, which thousands of inexperienced editors and/or vandals have inserted and continue to insert into articles without having any effect on its designation as non-RS. JoelleJay (talk) 16:17, 20 March 2023 (UTC)[reply]
Support The general point that sources might be in some format other than text seems sensible. For example, in the case of the recent train crash in Greece, a news report was cited and this was a video, not a text article. A reporter explained some details of the signalling system and it did not require OR to understand what was said. Another example would be articles about elections and places which will tend to use tables as sources to report numbers of votes, population sizes and the like.
As for maps, these are bound to be used for geographical topics and common sense applies. For example, see the current FA – Branford Steam Railroad. This includes a map in its infobox which seems to be drawn from Open Street Map. I'm not sure of the technical details but it seems sensible to provide this information in good faith to the reader, just as we might provide photographs of its scenery and features.
Support – I think it's time for WP:OR to explicitly address the use of maps. Maps are sometimes necessary sources for articles about things other than roads, like place names, but these citations can catch an editor by surprise if they're unused to reading or editing road-related articles. I think that's where the addition would have the most positive impact compared to the status quo. "Routine interpretation" in my view is a very literal interpretation. The most I've ever stretched WP:CALC was where I had to group rows in a poorly laid out spreadsheet. In that case, I felt it was necessary for the citation to include step-by-step instructions on reproducing these calculations (using a pivot table), because not everyone is comfortable using their preferred spreadsheet software in that manner. Similarly, a map citation would need to include more detailed instructions as the method of interpretation becomes more ambiguous or error-prone. So if a citation ends up looking like a term paper, that's when you know it's OR and no carefully worded exception in WP:OR would change that fact. – Minh Nguyễn💬01:11, 20 March 2023 (UTC)[reply]
Support per many others above. It is definitely possible to misuse maps as sources, but it's just as possible to misuse books. That's why we've have policies like OR. Ed[talk][majestic titan]01:46, 20 March 2023 (UTC)[reply]
Support—We should not make black-and-white determinations of whether a source is original research by its form. There is enough editorial review capacity available to point out and resolve inappropriate interpretation of sources in specific articles and among individual editors or groups of editors. VC 02:09, 20 March 2023 (UTC)[reply]
Support, and I'd be happy to see "databases" added to the list. Of course, just because non-textual sources are not banned doesn't mean that any given non-textual source is reliable for a particular claim. Textual and non-textual sources should be evaluated on their individual merits, for the specific context. WhatamIdoing (talk) 03:06, 20 March 2023 (UTC)[reply]
Oppose Per colleagues above. I think out current guidelines on Original Research and Reliable Sources already handle this. This is a broad carve out (any form of information, such as maps, charts, graphs, and tables) when editors in these discussions mainly speak about maps for roads. I have no idea what Routine interpretation means, and why it should only apply to non-text media. If routine interpretation is to mean the same way we read and interpolate text sources, then we already have adequate guidelines to cover that. If routine interpretation is to mean something like the example in the proposal below (verifying statements about ground cover), then it is original interpretation and should not be allowed. Vladimir.copic (talk) 05:53, 20 March 2023 (UTC)[reply]
Support the idea, the precise wording can bear some more discussion. In particular, the word "interpretation" unfairly makes it sound like a violation of NOR, but it is only intended in the same sense that reading of a text involves the "interpretation" of inferring meaning from sequences of words. Using the plain meaning of a map is the same as using the plain meaning of a text. On a side issue, some responses tried to identify "primary" maps, but not even the official maps of government survey departments are primary sources. They are expert secondary compilations of information from masses of primary data such as aerial photographs, satellite radar data, physical inspection, etc.. It is not essentially different from the way a historian writes a book based on the primary documents found in archives. Zerotalk09:25, 20 March 2023 (UTC)[reply]
This is true except when a dynamic map is generated automatically from a database of GIS objects. No human secondary analysis occurs for any single map location, depth, or layer, and so sites like Google Maps are primary in that sense. Someone being responsible for the code that renders billions of data points into a readable format is not the same as someone deliberately choosing which of those points to represent and personally validating them. JoelleJay (talk) 16:05, 20 March 2023 (UTC)[reply]
Support, though the wording needs to be tweaked slightly in order to be grammatically correct. As written, the second part of the proviso means that the techniques must be a meaningful reflection of the sources. Indeed, I'm not convinced that any mention of the techniques is necessary; I would suggest provided that there is consensus among editors that it [the interpretation] is a meaningful reflection of the sources. Rosbif73 (talk) 10:52, 20 March 2023 (UTC)[reply]
Oppose while reading a map is not original research, there are genuine issues with maps being used for original research which this proposal will make worse. I remember an article about a road which cited the date of the road's construction to two maps. The road was in one of them but not the other, and the article was using that to conclude it must have been built in between the dates on the maps. Right now that's a clear WP:SYNTH violation, but if this were to go in to the policy you could argue that it's a meaningful reflection of the sources and must be fine. Hut 8.518:42, 20 March 2023 (UTC)[reply]
We should not be using maps to conclude that something was built in the period between two consecutive maps. However, how would you feel if it were "The [something] first appeared on the [x] edition of the annual map"? Do you have any suggestion as to how the proposal should be worded in order to allow but not advocate the use of maps, in line with WP:RS? - Floydianτ¢20:42, 20 March 2023 (UTC)[reply]
Support - for obvious observable features only. An example, we can say that Mustang, Oklahoma is entirely within the boundaries of Oklahoma City, with a citation to a map, or to perhaps to use cardinal directions to say where something is located in relationship to another location. That said, there are concerns we should clarify before any final wording. We should not use a map to count the number of parks or religious structures in Mustang, or even use a map to determine distance, because that involves either the features of the map or subjective judgment.
Maps cited in articles should follow the Wikipedia:Reliable sources guidelines. When dynamic map applications (such as Google Maps or an ArcGIS website) are referenced, and when supported by the application, the URL used in the citation should link directly to an overview of the relevant object(s) rather than to the main page of the application.
Maps should only be used to cite statements that can be verified using the map. For example, contour lines on a topographic map can be used to reference statements about topography, but a statement such as "Washington D.C. is the most populous city on the Potomac" cannot be fully verified using a map that does not contain any information about population.
Proposal 2a: comments
Support—again, we allow foreign language sources to be translated by editors and used as source material. This would allow a visual source to be similarly translated, for what it visualizes, and cited as a source in an article. Basic map-reading skills are part of elementary school-level curricula in the United States, and those skills are probably taught at a similar age level in other countries. Thus, the skills needed to translate these sources may be more universal than that needed to translate foreign language texts. Imzadi 1979→06:27, 19 March 2023 (UTC)[reply]
Support As I stated above - not every map is reliable, just like not every newspaper is reliable or every book is reliable. But some certainly are. Maps should not be treated differently in this regard, either way. --Rschen775407:42, 19 March 2023 (UTC)[reply]
Strong oppose the use of dynamic map applications as a WP:V violation; static links must be provided to allow verification of the information, and if static links cannot be provided then the source is inappropriate to use. BilledMammal (talk) 11:18, 19 March 2023 (UTC)[reply]
The access-date is the static information. If the dynamic map can't verify the data sourced to it, then the article is out of date. - Floydianτ¢14:49, 19 March 2023 (UTC)[reply]
About if static links cannot be provided then the source is inappropriate to use: This is not an existing rule for any other type of source. What's special about a map website that makes you want to impose this rule on them alone? WhatamIdoing (talk) 04:46, 20 March 2023 (UTC)[reply]
Oppose – I don’t think the use of dynamic mapping services as sources should be encouraged (or even allowed). Their content changes unpredictably, and if I’m not mistaken, there’s no way to link to a specific place at a specific time on Google Maps – only a specific place at the most recent available time. That makes them inherently unreliable. I’m happy to support if the sentence about dynamic maps is removed though, since maps published in atlases etc. don’t have that issue. Justin Kunimune (talk) 12:51, 19 March 2023 (UTC)[reply]
I think there's an mis-definition here. "Dynamic" in the context of maps means the scale is adjustable, not that the content changes, per se. When linking to Google Maps, you can link to a specific location at a specific scale. The Michigan Department of Transportation has a dynamic mapping application called the Next Generation Physical Reference Finder, and it has fixed data sets within it. It's dynamic because you can also zoom in and out to adjust the scale being viewed. In other words, these are the opposite of fixed-scale paper maps or the digital copies of such fixed-scale maps. Imzadi 1979→15:33, 19 March 2023 (UTC)[reply]
Support. MLA Handbook for Writers of Research Papers (7th ed., 2009) specifically endorses Google Earth. Also as a person who trains for search and rescue with the Civil Air Patrol, both on the ground and in the air, I believe the use of paper maps is diminishing so rapidly that there is no alternative to dynamic online maps. Jc3s5h (talk) 14:11, 19 March 2023 (UTC)[reply]
Support - Maps are generally reliable in showing information, whether is is a paper map/atlas or an online map such as Google Maps. Dough487214:26, 19 March 2023 (UTC)[reply]
Support - A dynamic map is cited with an access-date. If someone looks at the source and finds it contentious, they can tag it as such. The article is cited to the map as it appeared on the access-date, much as a website is cited as it appeared on the access-date. Every website is dynamic. - Floydianτ¢14:57, 19 March 2023 (UTC)[reply]
Support per Floydian. I just wanted to point out that ArcGIS data sets are not necessarily dynamic but how they are displayed is. Often times, the data will have a version date. –Fredddie™15:21, 19 March 2023 (UTC)[reply]
Oppose, this actually appears to be an end run around OR by using a broad statement that actually includes most OR. For example "For example, contour lines on a topographic map can be used to reference statements about topography" appears to oversell how it could be used, 90% of statements about topography are too complicated to source to a topographic map... Those actually are OR. Horse Eye's Back (talk) 16:33, 19 March 2023 (UTC)[reply]
Topographic map reading is course taught in schools for a few different programs ranging from civil engineering, surveying, to wilderness survival. I understand the concern, yes this technique can be misused. Any source can be misused. However, if someone is using techniques taught in school courses on the subject, I don't see how that is OR. Dave (talk) 17:14, 19 March 2023 (UTC)[reply]
I am familiar with how to read a topographic map, I am also familiar with soil maps, GIS, nautical charts, etc. Doesn't all original research use techniques taught in school courses on the subject? Horse Eye's Back (talk) 17:20, 19 March 2023 (UTC)[reply]
So is the solution to ban books? Or publish guidelines on the proper use of books? The problem isn't the source, it's the misuse of the source. That is the same for books or maps. Dave (talk) 19:08, 19 March 2023 (UTC)[reply]
The solution you are proposing is to change the definition of OR entirely... "if someone is using techniques taught in school courses on the subject, I don't see how that is OR" would be a completely different conception of OR than we've used for the past 20 years or so. It would legitimize almost everything which is currently categorized as original research. Horse Eye's Back (talk) 19:17, 19 March 2023 (UTC)[reply]
@Horse Eye's Back, I'm not sure that I understand your statement that 90% of statements about topography are too complicated to source to a topographic map. Do you mean 90% of the statements that some editor might (incorrectly) claim is on the map, or 90% of the things that you could (correctly) identify from such a map yourself? WhatamIdoing (talk) 04:51, 20 March 2023 (UTC)[reply]
Do you think that 90% of the statements an editor would like to put in an article, and would claim are supported by a topographical map, would be wrong? WhatamIdoing (talk) 15:36, 20 March 2023 (UTC)[reply]
That first one is more blue sky and most likely doesn't need a source at all. That second one seem to be a personal judgement on what is steep or not, a statement like "the northern section of Lombard Street has a 35% grade" would be ok but the personal analysis isn't. Horse Eye's Back (talk) 15:40, 20 March 2023 (UTC)[reply]
A 35% grade would normally be described, in simple English, as "steep".
You have said above that 90% of statements about topography are too complicated to source to a topographic map. Can you give me some examples of some statements about topography that you believe are too complicated to source to a topographic map? WhatamIdoing (talk) 15:43, 20 March 2023 (UTC)[reply]
Oppose "Washington is the most populous city on the potomac" is, ironically, the only part of these two paragraphs that I could cite to a map while complying with WP:V. Citing your personal interpretation of contour lines is a good way to commit WP:OR, and citing to a website where the available data changes without notice and is then lost forever, like google maps, is a terrible idea. Meanwhile population size can be cited to most large atlases without issue. I'm quite confident something like the TIMES atlas of the world lists population sizes for major cities plus original source and date of census, and would be admissible as an RS. --Licks-rocks (talk) 17:33, 19 March 2023 (UTC)[reply]
Oppose the inclusion of Google Maps. As mentioned by other users, Google Maps is too ephemeral to use as a lasting citation. It is also unreliable – I frequently see errors on Google Maps, including the locations of towns and the paths of roads, and it has a lot of user-generated content. I've found it especially unreliable outside the US. —Mx. Granger (talk·contribs) 18:14, 19 March 2023 (UTC)[reply]
Then can we either provide an example that's more reliable, or omit the example? If we put the proposed text in a policy page, editors will inevitably interpret it as an endorsement of Google Maps's reliability. —Mx. Granger (talk·contribs) 19:05, 19 March 2023 (UTC)[reply]
What would you suggest? I think that if we just said "dynamic map applications" that would result in confusion given some of the comments at the bottom. Rschen775423:11, 19 March 2023 (UTC)[reply]
I don't have an example in mind. It should be something that's reliable, not based on inadequately vetted user-generated content, and has enough stability (or can be archived) so that one editor can add it as a source and another editor can check that source months or years later. If we can't think of a dynamic map application for which that's true, then the policy page probably shouldn't encourage editors to cite dynamic map applications as sources. —Mx. Granger (talk·contribs) 03:16, 20 March 2023 (UTC)[reply]
Meh Google Maps is neither as good or as bad as people are claiming, above. However, it has a strong western bias. I would rather people use a country's equivalent of a quarter quadrangle or a yearly tourist map first --GuerilleroParlez Moi19:18, 19 March 2023 (UTC)[reply]
Indeed. There are pros and cons - I have found that Google Maps is not as reliable outside the West, however it is more likely to be updated quickly after something changes, and there is the accessibility factor to editors. I am in the process of buying foreign road atlases, but some are just not available in the US (i.e. Brazil) - not to mention the financial commitment, including international shipping. What I don't want to do is put in a preference, because a preference turns into a black-and-white matter. --Rschen775419:57, 19 March 2023 (UTC)[reply]
Comment Dynamic map services can be reliable but not always easily citable. Some dynamic map services (like Google Maps) do provide a way to get a URL to return you to a specific view. Others don't and you can only cite the entrypoint URL, but so long as the coords are present, it is usually easy to go to the correct area and load the layer that supplies the relevant information (e.g. contours, watercourses, etc), but that does assume you have some familiarity with the layers that the map service uses (there are over 1000 layers in the dynamic map service my government provides for my state ranging from school catchments to surface mining leases). Certainly the template Cite map seems to be intended for maps published in books and not dynamic map services. I think having a template for dynamic maps (generally and/or specifically) with parameters for coords, layers, etc would help both the contributor and the reader when citing dynamic maps. Kerry (talk) 22:15, 19 March 2023 (UTC)[reply]
Strong Oppose. I don't necessarily object to something much stricter than this. But interpreting topography from contour lines, for example, is a clear example of OR as we have always understood it. What this is describing is vague and open to huge abuse. And besides that you can't use several of these kinds of online mapping services as citations because they do not have a permanently-citable form. Kahastoktalk22:30, 19 March 2023 (UTC)[reply]
Comment – Regarding the first paragraph, I support the use of deeplinks/permalinks. The issue with Google Maps is not so much that it's dynamic but rather that its editorial practices perhaps aren't as transparent as folks would like. For example, I can only prove circumstantially that Google Maps in Vietnamese uses Wikidata labels verbatim on its labels for non-major cities, without any quality control. (Compare e.g. [8] with this edit.) Yet it can be shown (again circumstantially) that this is not the case with its English labels. Similarly, Google Maps has incorporated a significant amount of user-generated POI data through its Map Maker and Local Guides programs, but it isn't isolated to a particular geography. If we can't reasonably distinguish the raw user-generated content from the vetted content, then citing Google Maps indiscriminately is rather like citing a news aggregator or a news organization that doesn't clearly label its opinion editorials. This limitation is not shared by every dynamic map service.
The second paragraph espouses a sound principle, but it could include a better example. A more direct example would be to say that "Washington, D.C. has a population of ####" based on the label "Washington (pop. ####)" on the map. At least in North America, many conventional paper maps include elevation, population, or other captions in place labels, or symbolize places differently based on size categories.
Regardless, I agree with pburka that this proposal on the whole constitutes advice on reliable sourcing, not drawing a line about original research.
Oppose, while the concept that sources should be used to cite what they say is correct, the examples show how tricky this is. Editors should not be making broad statements about topography based on contour lines. Perhaps identifying high and low points, but going further necessitates more personal interpretation. On the other hand, describing population from a map could be trivial, or it could be interpretation, depending heavily on what exactly is being said and what is on the map. CMD (talk) 02:07, 20 March 2023 (UTC)[reply]
Oppose: would this proposal advise editors to link out to maps without the equivalent of a permanent link? That's asking for trouble as we're building Wikipedia for the long term. Buildings, roads, and even some geographical features will change. We need to be able to verify those changes and other information now and in the future, whether it's to check citation integrity or write about a historical event. Unfortunately, we do not currently have a way to systematically archive map links/figure out which ones are functionally dead for our purposes like InternetArchiveBot. Using access dates, as one editor suggested above, instead of solving the fundamental verification issue here merely sets us up for a major future link rot problem. If I'm missing something here, e.g. if we are able to archive stable URLs that show what a map looked like at a specific time, please let me know. Ed[talk][majestic titan]02:14, 20 March 2023 (UTC)[reply]
Is there a policy that says that sources used must be able to be archived? I feel that this is a requirement much stricter than policy requires. --Rschen775402:22, 20 March 2023 (UTC)[reply]
See e.g. RoySmith's and Licks-rocks's words in #General comments as well. Not being able to verify this information, theoretically* as soon as the edit is saved, is a fundamental problem that goes against the spirit of WP:V. I'm not aware of any policies require archivability, Rschen7754, but I wonder if that's because we haven't really contemplated this. (Wikipedia:Link rot doesn't have anything to say about such a scenario!) *Theoretically though not practically.Ed[talk][majestic titan]02:34, 20 March 2023 (UTC)[reply]
Support—Maps should follow WP:RS just like other sources and should only be used to cite statements that can be verified using the map. I have no reservations about non-dynamic maps that are used correctly. However, the dynamic maps part is the weakest part of the proposal, and we should explore further how they should be used and cited after this RfC is over. VC 02:18, 20 March 2023 (UTC)[reply]
Oppose The issues of sticking to what the source says and allowing for changes to the source apply to any kind of source, not just maps. The proposed addition seems excessive detail contrary to WP:CREEP. Andrew🐉(talk) 10:00, 20 March 2023 (UTC)[reply]
Oppose as unnecessary; the wording in Proposal 1 plus the existing policies and guidelines are sufficient. Giving examples focuses too much attention on the examples themselves rather than on the principles they are trying to embody. Rosbif73 (talk) 11:02, 20 March 2023 (UTC)[reply]
Oppose There are multiple aspects to this, Maps cited in articles should follow the Wikipedia:Reliable sources guidelines. is completely unnecessary as everything used as a source should follow the reliable source guidelines. When dynamic map applications [..] are referenced [...] the URL used in the citation should link directly to an overview of the relevant object(s) rather than to the main page of the application. this is just a best practice guideline and is already covered by general guidelines to link to the specific part of a source where relevant and possible. Maps should only be used to cite statements that can be verified using the map. is redundant to general sourcing practices - if a map is used as a source for something it doesn't show then relevant statement has failed verification in exactly the same way as if it had been a textual source. Thryduulf (talk) 11:16, 20 March 2023 (UTC)[reply]
Oppose including this text. I agree with Thryduulf on this. It would help if Proposal 1 was more clear that all of the rules about citing text sources apply equally to citing maps. We shouldn't be giving the impression that texts and maps have different rules. Zerotalk12:28, 20 March 2023 (UTC)[reply]
Oppose as scope creep. As covered in WP:SPS - "Exercise caution when using such sources: if the information in question is suitable for inclusion, someone else will probably have published it in independent, reliable sources." and referenced in User:Uncle G/On notability. So you can use Google Maps to say "Charlie's Hard Brexit And Chips Shop is located at 123 Farage Street, Boston, Lincs" but it no book or newspaper sources mention the location anywhere, we should wonder why it's important to mention it in a general purpose encyclopedia. Ritchie333(talk)(cont)13:42, 20 March 2023 (UTC)[reply]
Proposal 2b: image layers
Satellite layers (i.e. on Google Maps) can be used to reference statements about elements such as ground cover that can easily be verified.
Oppose While I can certainly understand the rationale behind my fellow road editors using this and why they think this is uncontroversial, I personally find that interpreting the satellite layer is a bit more interpretation than I am comfortable with and I do not use this in my articles. --Rschen775407:42, 19 March 2023 (UTC)[reply]
It doesn't matter if I can or cannot; satellite imagery is a primary source, and per WP:PRIMARY any interpretation of a primary source requires a reliable secondary source. I also think you underestimate the complexities of interpreting satellite imagery; while you might successfully interpret simple features most of the time, I doubt you will do so 100% of the time. BilledMammal (talk) 15:30, 19 March 2023 (UTC)[reply]
As a GIS specialist, I can tell you that's harder than it seems in most cases. Grass, trees concrete? maybe. Grass, threes, concrete, granite, shrubs? starting to get difficult. Grass, savanna, peat, forest, plantation, burn scar? Impossible without advanced tools. --Licks-rocks (talk) 16:21, 19 March 2023 (UTC)[reply]
I think you're trying to squeeze the juice out of rocks here. I would use general terms like grassland, prairie, or pasture. I would certainly only mention specific species of grasses if they were talked about in a news story about prairie restoration. Note: I'm not relying on a map for that. If it's a case where someone is relying on a map for specific types of grasses, then we need to educate the editor, not throw out the map. –Fredddie™17:11, 20 March 2023 (UTC)[reply]
Not sure that works any better, how do you tell the difference been a grassland, a prairie, and a pasture from a satellite image? Those are all different things, not general terms for the same thing. Horse Eye's Back (talk) 17:16, 20 March 2023 (UTC)[reply]
Oppose I think there are statements about things like ground cover that can be objectively deduced from satellite layers, but I expect the two main use cases of this proposed policy would be biome classification and deforestation characterization, and I feel like satellite images can be misleading for both of those. We don’t typically cite bare photographs, and for the same reasons we shouldn’t rely on satellite images. In addition, Google Maps has the issues Billed Mammal and I mentioned above. Justin Kunimune (talk) 13:00, 19 March 2023 (UTC)[reply]
Oppose the use of satellite (and/or aerial) images as RS. They are simply too subjective, OR. Google street view does have dates taken, but still has the same problem. Use images to illustrate other RS but not RS by itself. Sammy D III (talk) 14:03, 19 March 2023 (UTC)[reply]
Support - Satellite imagery in a map visualizes what an area looks like, whether it is farmland, woodland, or development. It is basic interpretation to describe what an area is like by looking at satellite imagery. Dough487214:30, 19 March 2023 (UTC)[reply]
Weak support if using general terminology like Dough4872 suggested. I certainly wouldn't use it to tell what kind of crops are being grown in the farmland, though, being from Iowa, I could probably tell. –Fredddie™15:36, 19 March 2023 (UTC)[reply]
STRONG Oppose this would violate WP:OR, interpreting satellite data is the domain of cartographers and data scientists, not Wikipedians. --Licks-rocks (talk) 15:56, 19 March 2023 (UTC)[reply]
Strong Oppose LU/LC classification is more of an art than a science in my decade of experience as a GIS Analyst. I feel like it is solely original research based on a primary source --GuerilleroParlez Moi16:09, 19 March 2023 (UTC)[reply]
Strong Oppose, I work with literally the best machines learning tools for discerning ground cover available to civilians... Even that is just a guess, someone eyeballing a google earth image and then telling you what species of ground cover are in the image is insane to me. Even distinguishing between between broad categories say like a woodland and a tree plantation is nearly impossible. There is no easy verification here, the only way to do that would be to get boots on the ground (which again would be OR). Horse Eye's Back (talk) 16:22, 19 March 2023 (UTC)[reply]
Sometimes it's hard to even distinguish asphalt from water. Black Lake near Milford MI is not a lake. It's 67 acres of asphalt at the GM Proving Grounds. And, in the right light, it looks like water in aerial or satellite photos. Even ducks occasionally get confused and try to land on it. So, I Oppose this one. Banks Irk (talk) 16:58, 19 March 2023 (UTC)[reply]
I think your general point is valid, but not this specific example. A "real" lake would be explicitly labeled as a lake and colored blue on any political/topographical/recreational maps or map layers by the cartographer who is more qualified than the average Wikipedian. I absolutely agree that anybody who relies on a satellite image alone or ads a description based on a satellite image that is not supported or contradicted by say a political maps, should receive such an explanation of the pitfalls. Dave (talk) 17:26, 19 March 2023 (UTC)[reply]
Oppose. I don't see how this could be done without OR. It's also worth noting that the satellite layer on Google Maps isn't always correctly aligned with the map layer, especially in remote areas, areas with unusual topography, and areas that use nonstandard coordinate systems. —Mx. Granger (talk·contribs) 18:21, 19 March 2023 (UTC)[reply]
Isn't this throwing the baby out with the bathwater? The layers might not be aligned in spots therefore the whole thing is garbage? –Fredddie™19:39, 20 March 2023 (UTC)[reply]
Support however there should be some guidelines. I am totally fine with saying an area is "farmland" based on a satellite image review. That to me is approaching WP:SKYISBLUE levels of obvious. However, only "high level" summaries of the area are ok. As others have pointed out, the average Wikipedian would not be qualified to call something a sandy desert, or a pine forest based solely on a satellite image. If it is indeed a pine forest or sandy desert, those details would be more likely to be explicitly stated on a recreational map, which would be a better choice to use. Dave (talk) 20:15, 19 March 2023 (UTC)[reply]
How does one tell the difference between farmland, hunting land, natural meadow, dry seasonal wetland, marginal land, conservation land, private estates, research centers, etc from a satellite image? Thats actually way harder than seeing whether a desert is sand or gravel. Horse Eye's Back (talk) 20:22, 19 March 2023 (UTC)[reply]
This can easily go down rabbit holes and we can be playing what ifs all night. However just to answer the questions, In addition to what looks like manicured land in perfect circles or squares that could in no way be natural, farmland would have supporting evidence visible such as presence of tractors, combines, barns, silos, irrigation systems, etc. that would be missing if it were say conversation land. If supporting evidence were lacking, then I wouldn't feel comfortable with calling it farmland. Dave (talk) 20:34, 19 March 2023 (UTC)[reply]
So just pure OR then? All of that could be found on conservation land that was formerly farmland or at a museum or managed hunting land. Also note that much farmland is not in circles, rectangles, or squares... It might not even be visibly cleared land as in the case of tree farms... It might not look like manicured land at all. Horse Eye's Back (talk) 20:43, 19 March 2023 (UTC)[reply]
Oppose. This is blatant SYNTH/OR. Unlike the info extracted from maps being championed by supporters, this proposal is actually SKYISBLUE-level obvious in its non-compliance with policy. JoelleJay (talk) 23:34, 19 March 2023 (UTC)[reply]
Support. If an editor cannot be relied on to look at a picture of a forest and call it a forest, they have no business editing any sort of nonfiction work, much less an encyclopedia. —Scott5114↗[EXACT CHANGE ONLY]23:58, 19 March 2023 (UTC)[reply]
Comment – I feel like I'm missing some backstory behind this proposal. What is the purpose of characterizing landuse or landcover from satellite imagery in an article? It seems oddly specific for a project-wide guideline. I do think it would be helpful to clarify that "maps" doesn't necessarily mean four-color road atlases and can include imagery, but we probably already have a guideline somewhere about citing an image for its self-evident contents, no? If we're going to carve out a safe harbor here, a better one would be street-level imagery, which anyways can be a much more reliable tool for determining LU/LC when it's available. – Minh Nguyễn💬01:57, 20 March 2023 (UTC)[reply]
Oppose—Satellite layers do not have an editorial process related to the features that are depicted. The task of creating a legend to clearly define what a particular shade of green means or what that thing that looks like crosshatching means is nearly impossible. Therefore, satellite layers require too much expert interpretation or speculation about what is shown to make me comfortable about them being used as sources. VC 02:30, 20 March 2023 (UTC)[reply]
Oppose as both unnecessary and not always correct. Proposal 1 allows for satellite imagery to be used when appropriate (e.g. "2018 satellite imagery showed the area to be forested, but by 2021 suburban development was apparently complete.") but interpretation will sometimes be original research. Thryduulf (talk) 11:24, 20 March 2023 (UTC)[reply]
It is allowable to cite historical maps to refer to how artificial geographical features appeared on them at a specific point in time.
Proposal 3: comments
Oppose: I think that this runs contrary to WP:OR as it does not just come from the older source but explicitly requires the comparison between the older source and a newer source. Gusfriend (talk) 06:15, 19 March 2023 (UTC)[reply]
Support—we cite historical documents and works all the time for the representations of information known at that time. Allowing a map to be cited but not allowing a historical map to be cited for the information presented on it would question the ability to cite any historical work of any kind. Imzadi 1979→ 06:36, 19 March 2023 (UTC) Amending to add: failure to allow historical maps would mean, at a minimum, that a 2023 map could be used today to cite the description of the location and routing of a current highway/rail line/etc., but a 1939 map could not be used to cite the description of the location and routing of a highway/rail line/etc. that was taken out of service later that year. Imzadi 1979→06:50, 19 March 2023 (UTC)[reply]
Support It is not original research to state what is on a map published at a certain point in time. It is original research to read details into a map that are not there, just as with graphs and charts, and textual sources. Sure, it is preferable to use newspapers because maps cannot convey the reason why events happened as they did, and because dates given can be more precise. But in many cases, there are years of archives missing from newspapers, if there are online archives at all. --Rschen775407:43, 19 March 2023 (UTC)[reply]
Oppose. As Gusfriend points out, this requires comparing two sources and drawing original conclusions, a clear WP:OR violation. There are also multiple explanations for why two maps may provide different information; changing features is one possible explanation, but others are that the features are trap streets, or that one or both of the maps were incorrect at the time of production. Determining which explanation is the correct one requires considerable WP:OR or WP:SYNTH; instead, we need to rely on explanations from reliable sources. BilledMammal (talk) 11:12, 19 March 2023 (UTC)[reply]
Support I don’t think this is any worse than using other kinds of historical sources. If someone says something like "feature X was present on this official 1813 town map but is absent on this local roadmap in 2016", that’s valid and useful information. There might be any number of confounding factors, but if the maps come from reliable sources, I would think a change in the feature itself is the most likely explanation. Justin Kunimune (talk) 13:12, 19 March 2023 (UTC)[reply]
Even reliable sources included trap streets, so being reliable doesn't guarantee that the feature existed. It also doesn't guarantee that the source didn't make a mistake, and we risk propagating that mistake if we rely on the difference between a historic source and a current source to say that a change occurred, rather than that an error was corrected.
I also note that per WP:SYNTH we can't do this with other kinds of historical sources either; if a book from 1813 says that "disease spreads by miasma" and a book from 2016 says that "disease spreads through germs" we cannot draw the conclusion that sometime between 1813 and 2016 disease stopped spreading by miasma and starting spreading through germs unless a source says "in 1880 disease switched from spreading by miasma to spreading through germs", because to do otherwise would combine material from multiple sources to reach or imply a conclusion not explicitly stated by any source. BilledMammal (talk) 14:11, 19 March 2023 (UTC)[reply]
Even reliable sources included trap streets, so being reliable doesn't guarantee that the feature existed. Sounds like a contradiction to me. --Rschen775414:55, 19 March 2023 (UTC)[reply]
Possibly, but as trap streets were very common in maps if we were to consider maps that include them unreliable we would need to consider most maps unreliable. In addition, there is the practical issue of identifying which maps include trap streets - they rarely disclosed the fact. BilledMammal (talk) 15:15, 19 March 2023 (UTC)[reply]
Throwing out maps because there might be a trap street is like throwing out books because a word might be misspelled. Seems like a WP:BABY situation. –Fredddie™15:41, 19 March 2023 (UTC)[reply]
Wouldn't it be more equivalent to a book which randomly had false information inserted in an attempt to prevent copyright violation? In that case we certainly would treat the source as unreliable. Horse Eye's Back (talk) 17:24, 19 March 2023 (UTC)[reply]
Hypothetically, if an author said he or she inserted a copyright trap into a book and didn't say where it was, the whole thing would be unreliable? What if the author said where the copyright trap was? Would that validate the whole minus the copyright trap? –Fredddie™18:58, 19 March 2023 (UTC)[reply]
Then it wouldn't be copyright trap would it? It would be an acknowledged error and we would handle it as such. I would note that if you try to submit academic papers to reputable journals with copyright traps you're either going to be denied publication or the paper will have to be retracted when the trap is discovered. Its not considered acceptable behavior in academia, its generally considered the same as knowingly publishing false information (which for us would mean deprecation). Horse Eye's Back (talk) 19:39, 19 March 2023 (UTC)[reply]
University textbooks, published encyclopedias, even telephone directories have contained copyright traps. We don't discard the Encyclopedia Britannica because it contained copyright traps. I don't see how that's a relevent argument. Dave (talk) 19:28, 19 March 2023 (UTC)[reply]
The Encyclopedia Britannica contains copyright traps? I can't find anything about that on google and it hasn't been discussed at RSN, what is your source? Horse Eye's Back (talk) 19:35, 19 March 2023 (UTC)[reply]
I think the larger issue is that such observations will always be SYNTH/OR because they do not have secondary sources pointing them out. That is never, ever acceptable on wikipedia. JoelleJay (talk) 23:27, 19 March 2023 (UTC)[reply]
Support - Maps are fine to describe how a road or place was at a certain point in time. It is probably a good idea to use two maps as a source to prove a change happened by a certain time. Dough487214:32, 19 March 2023 (UTC)[reply]
Oppose This is very much in OR territory, because it assumes all maps for a region from RSes have been found to make this evaluation. All you need is one missing map to make the assumption flawed. --Masem (t) 14:32, 19 March 2023 (UTC)[reply]
Its the comparison between two or more sources to come to an editor-made conclusion that something was first, last, or introduced or removed at a certain time which is not explicit in the sources. We're not supposed to be doing that anywhere. On the other hand, if we have a single source that spells that out, that's fine. Masem (t) 14:45, 19 March 2023 (UTC)[reply]
How about when it's spelled out in the article as "The [feature] appeared for the first time in the state's official map for 1947" or "By 1947, the [feature] appeared in official state maps."? - Floydianτ¢15:00, 19 March 2023 (UTC)[reply]
That to me is still pushing the OR. Arguably you should be staying something like "That [feature] first appeared as early as 1947" to leave open the possibility that a yet-unfound 1946 map may have it, and that can be rather weak writing. Masem (t) 15:05, 19 March 2023 (UTC)[reply]
And yet, if you have a known and library-cataloged corpus of annual maps that includes editions for 1919, 1920, 1921 ... 1945, 1946, 1947, 1948... 2022 with every expectation that 2023 is being published, then you don't have that yet-unfound 1946 map. If 1946 is missing from the catalogs, then that sentence may need to be worded differently to reflect the sources as you suggest. Imzadi 1979→16:44, 19 March 2023 (UTC)[reply]
Depends on what you do at the library, if you count the number of columns the library has and then add that information to the library's wikipedia page that is OR. Same goes for counting the columns in a picture of the library, even if that picture is taken from space. Horse Eye's Back (talk) 23:12, 19 March 2023 (UTC)[reply]
What about if you go to the map room of the library to look at a floor plan or an engineering plan of the library, published by the builders of the library, in which the legend indicates + as a column? Would that be OR, or is counting no longer a basic calculation? - Floydianτ¢23:21, 19 March 2023 (UTC)[reply]
Counting (turning data which isn't a number into a number through analysis) was never included in WP:CALC. If the source says that one hall has 12 columns and the other has 8 an acceptable use of WP:CALC would be to say that there are 20 columns in the two halls. If the source doesn't give us the numbers to do the calculations we can't do them, we don't get to do analysis and then calculations from that analysis... That is the very definition of OR. Horse Eye's Back (talk) 23:34, 19 March 2023 (UTC)[reply]
Being mentioned in secondary sources is important for notability, but that guideline is for deciding whether a Wikipedia article should exist for a topic. Notability does not apply to a sentence or a paragraph. Jc3s5h (talk) 01:21, 20 March 2023 (UTC)[reply]
And, for better or worse, no policy requires all articles to cite secondary sources, much less to use them exclusively. We use primary sources precisely to include information that isn't "noted" in secondary sources but which encyclopedia articles are expected to contain (e.g., birth years for living people). WhatamIdoing (talk) 05:02, 20 March 2023 (UTC)[reply]
Of course all articles have to use secondary sources, OR explicitly states this. Primary sources are only allowable for very straightforward statements of basic facts. Making novel observations comparing sources is definitely not among these exceptions. JoelleJay (talk) 05:33, 20 March 2023 (UTC)[reply]
Primary sources can only be used for straightforward, basic facts. It has nothing to do with notability and everything to do with OR: you cannot draw conclusions from primary sources that have not been drawn in secondary sources. JoelleJay (talk) 05:35, 20 March 2023 (UTC)[reply]
Yes, but you said that "If that info isn't noted in secondary sources it has no business being on wikipedia anyway", which sounds like you're saying primary sources can't be used at all. WhatamIdoing (talk) 15:48, 20 March 2023 (UTC)[reply]
I am responding to How about when it's spelled out in the article as "The [feature] appeared for the first time in the state's official map for 1947" or "By 1947, the [feature] appeared in official state maps."?. Even if that info is verifiable, if secondary sources are not commenting on that feature's appearance it is original research and improper WP:PROPORTION to include that observation in an article. JoelleJay (talk) 16:29, 20 March 2023 (UTC)[reply]
There are well-known collections of annual official maps. For instance, the Michigan Department of Transportation, and its predecessor the Michigan State Highway Department, has published an updated official state highway map every year since 1958 except 1959 and 1994. They had semi-annual dated map editions from 1919 through 1957. Paper copies of all of them are available through the Library of Michigan, and the Archives of Michigan is placing digital copies of them online. This is just one state DOT. I know off the top of my head that full archives of annual editions have been scanned and made available online for Minnesota, Wisconsin, Illinois and Ohio. Many other states have done the same, and others are continuing to come online all the time. If there is a question about a possible missing edition, the Road Map Collectors Association has cataloged known editions. Additionally, Rand McNally has published annual editions of its Road Atlas for decades, although there are considerations there as the 2024 edition has already been released. Imzadi 1979→15:42, 19 March 2023 (UTC)[reply]
In the context of covering highways in Michigan the publications of the Michigan Department of Transportation would be primary sources and as such any analysis is strictly forbidden and will remain forbidden no matter what the consensus in this discussion is. Horse Eye's Back (talk) 16:54, 19 March 2023 (UTC)[reply]
"Strictly forbidden"? That's a bizarre interpretation of policy. Near as I can tell the relevant passage of current policy for that issue is "A primary source may be used on Wikipedia only to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge." So yes, there are restrictions on use, but that isn't the same as strictly forbidden. Dave (talk) 22:17, 19 March 2023 (UTC)[reply]
A statement like "Highway H appeared in the 1973 edition but not in the 1972 edition" is not analysis but just reporting what is in the sources. An example of improper analysis would be "Highway H didn't exist before 1973" since even official maps can show features in advance or delayed from their physical existence. The OR rules were invented to exclude material that can't be verified in published reliable sources, not to make editors' work more onerous. Zerotalk09:41, 20 March 2023 (UTC)[reply]
But if secondary sources are not commenting on that observation, it is against WP:PROPORTION to include it in an article. Doing so is reporting the editor's comparison and evaluation of the sources, which is OR. While it might seem like a harmless fact, allowing that leeway would have significant consequences in how contentious topics and their sources are presented, e.g. geopolitical territorial disputes: Evidence that politically-charged edits have the potential to make it past Google’s moderators is offered by Geens (2012), who documented edits to Syrian highway names honoring opposition heroes and events in country’s current civil war..[9]JoelleJay (talk) 16:44, 20 March 2023 (UTC)[reply]
This objection makes no sense because the proposed text doesn't say anything about a comparison. The proposal is just to say that a particular feature does or does not appear on a particular map. This indicates how things were at that point in time. Expecting editors to find and cite every map which includes the feature is absurd because some places appear on thousands of maps. Making an exhaustive analysis of them would be more like OR than what the proposal suggests. Andrew🐉(talk) 10:35, 20 March 2023 (UTC)[reply]
Support. I would not use a map as definitive evidence of when something began to exist or ceased to exist, as maps can't be relied upon to accurately reflect the precise situation at their moment of publication; but that doesn't seem to me to be what the proposed wording above talks about. When adding geographical coordinates to articles, I have often used old Ordnance Survey maps to determine exactly where no-longer-existing buildings, railway stations, etc., in Great Britain were located, and that seems an admissable use to me. Deor (talk) 15:47, 19 March 2023 (UTC)[reply]
Oppose This would be WP:SYNTH. We should cite a map or other publication that summarises the development of information on older historical maps for this, not the maps themselves.--Licks-rocks (talk) 15:59, 19 March 2023 (UTC)[reply]
Oppose. This example may or may not be permissible depending on the context. Giving blanket permission to publish original research based on historical maps would be a mistake, and the guidelines already describe permissible uses of historical and primary sources. pburka (talk) 16:22, 19 March 2023 (UTC)[reply]
Comment: It's fine to use a map as a primary source for a claim like "The city of X is shown on a 1634 map." Is this the kind of claim that's intended in this proposal? Or is it supposed to allow claims that require more interpretation than that? —Mx. Granger (talk·contribs) 18:26, 19 March 2023 (UTC)[reply]
That is the baseline that would be established by this proposal. There is more ambiguity that would need to be resolved in the future, perhaps through another RFC. Rschen775418:49, 19 March 2023 (UTC)[reply]
Comment: I find this rather ambiguously stated; it's hard to me to tell what's being authorized here. In general I agree with Deor's comments. Historical maps can be useful, but one needs to be careful not to claim more than can be strictly justified. Choess (talk) 19:58, 19 March 2023 (UTC)[reply]
Support (nom) However there needs to guidelines. I have myself found Wikipedia articles that inappropriately use maps in this fashion and have advised editors why I think they made a mistake. However, it's also a valuable resource that if used correctly can improve an article. If used incorrectly the likely result is an incorrect date or anachronism in the article. That also is likely the result if a historical newspaper were cited, but interpreted in a modern context. I wouldn't ban using historical newspapers, but I would support guidelines that remind people of the pluses and perils of using them. Same for maps. Dave (talk) 20:25, 19 March 2023 (UTC)[reply]
Oppose. Again, as far as I'm concerned what is described is textbook OR. But, to be clear, this is not because the map is historical but because of how it is proposed to be used. Kahastoktalk22:32, 19 March 2023 (UTC)[reply]
Support Everything we cite in Wikipedia is historic due to the fact that it had be created before we could read it and cite it. Whether it is historic from yesterday or historic from 100 years ago, there is always the need to be clear about the time period you are referring to when writing on Wikipedia. In my state, we have had small towns be relocated to higher ground following a major flood. So their location has changed over time and only the historic map (prior to the flood) can give you the original location. Yes, there are skills in reading a historic map, just as there is a skill in reading and interpreting information in a historic book where terminology may have changed over time etc), or in interpreting a graph, a table of statistics, or content in a foreign language. I can "read" historic maps from my state quite well (I do it all the time); I can't read books written in Italian or interpret chemical formulae. Sometimes I encounter a historic map on which I cannot get any reliable reference points (e.g. no mountains or other features that are going to be persistent to the present day, no coords along the axes), but that is very rare (and usually would not be the case on government maps) and I would not use information from that map (as I could not understand it), any more than I would cite a written source where the print was unreadable, nor would I attempt to explain chemical formulae etc. Not all contributors have the same skills, but, so long as we have the skills to interpret sources relevant to the topic area, I don't see why that is a problem. I write about history and geography of my state so its historic maps are essential to much of what I write about. And I don't write about chemistry as I know too little about it to meaningfully interpret its source material. I don't believe that I should criticise the content written by those who write about chemistry just because I cannot interpret the sources that they can, nor vice versa. If we prevent people contributing to Wikipedia in areas where they have the expertise to interpret sources, what kind of encyclopedia are we creating? We want to attract people with the skills to write accurate content on complex topics, not drive them away. If you look at the 2011 profile of Wikipedia contributors wrt to age and education, you will see that 73% of contributors in the survey were aged over 22 (an age by which a person might have acquired a tertiary qualification) and indeed 61% of contributors had such a qualification, so, looking at that survey, Wikipedians are quite highly educated, so we should not be surprised that there are some high levels of skills in interpreting a range of source material amongst our contributors. We should be celebrating and making the best use of such skills. Kerry (talk) 23:06, 19 March 2023 (UTC)[reply]
Oppose. This is blatant SYNTH/OR. Unlike the info extracted from maps being championed by supporters, this proposal is actually SKYISBLUE-level obvious in its non-compliance with policy. JoelleJay (talk) 23:33, 19 March 2023 (UTC)[reply]
How would you suggest rewording the proposal to accommodate "the info extracted from maps being championed by supporters", which it seems you have an inkling of support towards? I for one don't support the extremes that have been raised, but there is certainly something between "Maps no bueno" and "I've calculated the population density of this area using the 1984 Shell Oil Map" - Floydianτ¢20:37, 20 March 2023 (UTC)[reply]
Support. In many cases, a map can more accurately reflect a change to a geographic area than can a text source describing it in prose. From my experience working with such sources, textual descriptions of changes over time are actually more likely to get an editor's wires crossed than a map. —Scott5114↗[EXACT CHANGE ONLY]00:36, 20 March 2023 (UTC)[reply]
Comment, like some previous commentators I am not exactly sure what this is trying to say, and how it would be interpreted. I suppose this is thus an oppose on this wording, but I don't know if there is another wording I could support. CMD (talk) 02:16, 20 March 2023 (UTC)[reply]
Support—Stating what a map shows for a particular cited date is not original research. Interpretations involving multiple maps require greater scrutiny, but they should not be outright banned. VC 02:37, 20 March 2023 (UTC)[reply]
I've never opposed the use of maps. I don't believe JoelleJay has either. As far as I am aware nobody is opposed to the use of maps. All of these conversations are about which maps to use and how to use them not whether or not to use maps lol. Horse Eye's Back (talk) 20:57, 20 March 2023 (UTC)[reply]
Support. As long as claims are worded to say that this is how a given feature was depicted, which may or may not be how it actually was on the ground then this is fine. Thryduulf (talk) 11:27, 20 March 2023 (UTC)[reply]
Support this literal statement, but it isn't really necessary as current policy already allows us to say what is depicted on a map. There is a big difference between "map M shows X" and "X existed then". The first is just a report of a source, while the second requires some assumptions. It is also perfect fine to say "map M1 shows X1, while map M2 shows X2". This is just a report of two sources and we are allowed to report on as many sources as we like. However, writing "map M1 shows X1, while map M2 shows X2, and therefore Y" is likely to be a SYNTH violation depending on what Y is. All of this follows from existing policy and doesn't need special consideration for maps. Zerotalk12:23, 20 March 2023 (UTC)[reply]
Support per my John Roque's example above, but again with caveats. For example, you could say "The area east of Essex Road is marked as poor by Charles Booth's London Poverty Map of 1898" but simply saying "The area east of Essex Road was poor" would stray into original research, in my view. Ritchie333(talk)(cont)13:47, 20 March 2023 (UTC)[reply]
Having seen a "Good Article" that was exclusively referenced by about 30 maps with no non-map references it would be good to have a discussion about what maps mean for notability with a dip into the extent to which maps make quality references. Gusfriend (talk) 06:11, 19 March 2023 (UTC)[reply]
There is also the concern that sources such as Google Maps do not have a version and may be updated at any time which may limit what it is appropriate to reference. Gusfriend (talk) 06:19, 19 March 2023 (UTC)[reply]
The problem with Google Maps is that, while the basic map is reliable, some of the labels of various, usually man-made, features is done by a process of User Generated Content which is open to abuse. Murgatroyd49 (talk) 10:11, 19 March 2023 (UTC)[reply]
Doesn't this same consideration apply to any website? If this is an issue, any web cite should include an archived link, no? - Floydianτ¢15:02, 19 March 2023 (UTC)[reply]
Yes, that same consideration applies to any website, and always has. Archives may be nice, but it's more pointful to remember that if Google Maps says something one year but not the next, the information in the article is not out of date and needs to be fixed. Wikipedia isn't some WP:MMORPG where you win points for fancy citations. The important part is the facts reported in the sentence before the citation. WhatamIdoing (talk) 05:05, 20 March 2023 (UTC)[reply]
@Gusfriend: that is a good question, but that's not what is being asked. If you'd like, I'd suggest that you can start a discussion on that question later once this one is settled. Imzadi 1979→15:54, 19 March 2023 (UTC)[reply]
"Having seen a "Good Article" that was exclusively referenced by about 30 maps with no non-map references" - in my view such an article should be delisted and the article sent to AfD (or possibly redirected). Ritchie333(talk)(cont)13:56, 20 March 2023 (UTC)[reply]
That sounds like an over-reaction. It's not good to judge an article's notability on the basis of the sources someone has already cited. WhatamIdoing (talk) 15:50, 20 March 2023 (UTC)[reply]
Questions - are maps considered primary sources (with the restrictions and cautions of such), secondary sources or tertiary sources? Or some mix of all three? Does it depend on the specific map? Does it depend on the specific information WE are attempting to cite to the map?
A lot of the same questions could be asked for books and for newspaper articles. I cannot find where in our official guidelines we make these determinations. --Rschen775414:54, 19 March 2023 (UTC)[reply]
Assume maps are secondary, just as many assume news articles are secondary. Inclusion of a city on a state map would be just as much incidental coverage of that city as inclusion of the street address of a building is incidental coverage of that city street. Now a city map would be more significant coverage of that city just as a news article about a building is significant coverage of the building. In both cases, you have to evaluate the source to determine how it factors into notability concerns, but the questions above aren't about notability, but the ability to use a source. In other words, can you use that news article about the building in the first place? Can you use that map of the city or the state's highway system in the first place?
In other words, you have thoughts worthy of discussion, but they're not germane to the specific questions above and should be in a different discussion after the questions above are settled. Imzadi 1979→16:37, 19 March 2023 (UTC)[reply]
Its context that matters. For example a map of Pennsylvania highways published by Rand McNally is not primary for coverage of Pennsylvania highways (although it still isn't significant enough to count towards notability) but a map of Pennsylvania highways published published by PennDOT is primary in the context of coverage of Pennsylvania highways. Also note that a city map doesn't meet the significant coverage standard. Horse Eye's Back (talk) 16:42, 19 March 2023 (UTC)[reply]
You're confusing first-person vs. third-person authorship with primary vs. secondary sources. A map by the Maryland DOT shows most/all of the state of Delaware. MDOT's map is first-person for Maryland's highways but third-person for Delaware's. DelDOT's map would be third-person for the highways shown in Maryland. Both maps would have the same primary/secondary classification though. Rand McNally's respective maps in its atlas would both be third-person for both states, and yet share the same primary/secondary classification as the DOT maps. Since these otherwise equivalent maps are based on underlying source material (aerial surveys, ground surveys, photography/videography, etc.) and distilled together as a reporter distills interviews or reviews documents to craft news coverage subject to editorial oversight, if that news article is secondary, then perhaps that map is too. Imzadi 1979→16:58, 19 March 2023 (UTC)[reply]
IMO the primary/secondary classification is in f contextual. CNN is not a primary source for a story about a Chimpanzee sanctuary, they are a primary source for a story about CNN. I would also be wary of using margins, its a reliable source for what its about (a map of France is a reliable source for France) but information beyond the extent of the map would be a questionable use. Horse Eye's Back (talk) 17:12, 19 March 2023 (UTC)[reply]
Most maps are secondary or tertiary sources based on a mixture of primary sources — such as gps tracks, surveys, and imagery — as well as other maps. -- GuerilleroParlez Moi19:25, 19 March 2023 (UTC)[reply]
Pasting my argument from above: Unlike published paper atlases or road maps where the scale of a map and which details to include have been chosen deliberately and edited for accuracy by human authors, online database-generated maps with continuous spatial resolution have not received such oversight or secondary analysis at every zoom level. The latter group includes the GIS data-rendered dynamic maps from Google et al that are favored by highway projects and used to justify inclusion of arbitrary minutiae such as exact mile markers for individual localities and intersections along a route. Google Maps does not distinguish between cities and tiny non-notable communities (and so the choice for which ones to include on wiki will necessarily involve OR and likely circular referencing, with editors clicking on a Maps locale to bring up the linked wiki article and judging from its contents whether it should be mentioned), and these locations are not visible at every zoom depth; nor does it contain exact distance measurements between points (editors use various extensions for these calculations). For these reasons (not to mention known accuracy issues) the data extracted from such maps are not easily verifiable in the ways excepted by CALC nor are they DUE as they have not received secondary analysis. JoelleJay (talk) 23:46, 19 March 2023 (UTC)[reply]
I'm not opposed to using maps as sources, with some caveats. First, as others have pointed out, some (Google Maps, for example), are dynamic to the point where it's impossible (short of taking a screenshot) to cite a specific edition; that's a non-starter for me because with without a reproducible way to reference the material, WP:V can't happen. And, as others have also pointed out, it may not be obvious what information on the map is WP:UGC. And of course, there may simply be errors, but that's no different from any WP:RS. Stories like https://www.fastcompany.com/1700270/how-google-maps-led-accidental-invasion are not unheard of. Some maps intentionally show disputed borders, placenames, etc, differently depending on where the map is displayed. So, don't disallow all maps, but also don't blindly take them as gospel. -- RoySmith(talk)16:45, 19 March 2023 (UTC)[reply]
@RoySmith: Dynamic here is a term of art that means something closer to interactive than always changing. But, to your point Google Maps is always updating and it does do some wonky localization things -- GuerilleroParlez Moi19:22, 19 March 2023 (UTC)[reply]
Note that as a term of art Dynamic actually has at least three meanings in the context of maps, one is what RoySmith brings up, one is what you bring up, and there is a third which nobody appears to have mentioned at all which is actually how the term is most commonly used in the academic literature "Dynamic mapping is a cartographic concept used to depict dynamic spatial phenomena or to present spatial information in a dynamic way." [10]. Horse Eye's Back (talk) 19:29, 19 March 2023 (UTC)[reply]
The gist of my comment was that there's no way (that I know of) to create a URL which always gets you the same map data. I guess you would call that a permalink, but I'm not even sure that's the correct word (a wikipedia permalink, for example, still depends on the current value of transcluded templates). I should note that this is not unique to maps. For example, https://www.nytimes.com/2023/03/19/business/economy/fed-silicon-valley-bank.html will probably always get you to some version of the same article, but the version I'm looking at right now says, "March 19, 2023 Updated 3:42 p.m. ET". I don't know what changed 4 minutes ago, but something did. It might have been correcting a typo, or it might have been something significant. -- RoySmith(talk)19:49, 19 March 2023 (UTC)[reply]
I concur with several others that there seems to be a large WP:OR issue here. Maps are inherently hard to cite, and there need to be fail-safes included for that, with a detailed explanation for how to avoid verifiability and original research issues. In the current form of the RFC, every sentence proposed after the first seems to go further and further beyond the boundaries of WP:OR and WP:SYNTH, and there's very little proposed to rein these issues in. Citing a bird atlas for the known extent of the red bellied wood-piper or whatever is probably fine, but citing google earth for land use information is a terrible idea for so, so many reasons. If we're going include this, I think it should be made very clear that verifiability should be paramount. No citing dynamically updated material, no personal interpretation of lines on a map, and certainly no personal interpretation of satellite images. --Licks-rocks (talk) 17:09, 19 March 2023 (UTC)[reply]
On the topic of "personal interpretation of lines on a map", it's more than just lines. For example, while it might seem reasonable to look at a map and say, "Texas is larger than New Jersey", unless you have a certain level of understanding of map projections, you might be tempted to say, "Greenland is larger than South America". It can be tricky to tell what's obviously true and what's WP:OR. -- RoySmith(talk)17:31, 19 March 2023 (UTC)[reply]
In the roads space we've historically had a lot more OR than that, we're talking about OR like "Route 9 then heads east between a residential area and a car dealership" See for example Pennsylvania Route 309 or the route description of just about any American highway article. Horse Eye's Back (talk) 17:33, 19 March 2023 (UTC)[reply]
I'm broadly in favour of editors being allowed to cite selected maps as sources, but I'd want to nail down some specifics. For example, here in the UK, I'd say that the gold standard of reliability would be an Ordnance Survey map of 1:25,000 scale or greater. (The 1:1,250 scale maps are largely accurate too but they're so comprehensive as to be indiscriminate in what they include.) I would not necessarily say that a map in a printed book was reliable even if that book was otherwise a reliable source, because of the way publishing works. The book's text might have been written by Professor Verity Recondite of Sagacity College, Cambridge, but she might very well have had little to do with the maps. I would also want to say explicitly how appearing on a map interacts with notability. Here in the UK, something's likely notable if it's marked on an Ordnance Survey map of 1:100,000 scale or greater, but appearing on a 1:25,000 scale map is not evidence of notability. We'd need equivalent rules for you foreigners.—S MarshallT/C20:33, 19 March 2023 (UTC)[reply]
Wow, the size of this discussion points out one reason I seldom look into VP, but a Discord post alerted me. I'm not a roadie; rather I do a lot of buildings which have similar problems. Google's databases are less reliable for buildings than for roads, and I long ago gave up trusting them for this. Wikidata locations are perhaps even worse. So, what I do is use Google Earth aerial to look at the indicated location and then Street View for my target and, if that produces a doubt, do a websearch for a photo or physical description to compare to the aerial and street views. Aerial is highly reliable; the pictures are well aligned and the coords come out correct to ground truth by going there with my GPS phone. Google Street View is sometimes referenced to the wrong street or otherwise in error, but is highly precise when it doesn't have obvious errors. Streets, when they have errors in Google, are errors in the database and the pictures are again far more reliable. And no, I don't try to establish an article topic's notability by presence in Google's databases or Wikidata or GNS or other database. Notability requires serious discussion (words) published in a reliable source. Jim.henderson (talk) 20:55, 19 March 2023 (UTC)[reply]
I routinely use Google Earth for my job as a utility engineer. The UTM coordinates it spits out are often more reliable than the databases of infrastructure owners (i.e. a power pole from the power utility company's GIS database gets less consideration than a crisp up-to-date satellite image from Google Earth)! - Floydianτ¢23:17, 19 March 2023 (UTC)[reply]
Somewhat off-topic, but in case anyone reading this is wondering how it's possible to evaluate the accuracy of Google Maps or any other imagery provider: In my own contributions to OpenStreetMap (not nearly RS, hah!), I don't have access to Google Earth for IP reasons, but among the imagery I do have access to, I generally prefer government-provided orthoimagery layers. Often I'm able to track down the RfP for the imagery collection, which states a maximum horizontal tolerance that's plenty good enough by OSM standards. It's also possible to check the alignment of any orthoimagery by comparing markers to authoritative sensors, e.g. using [11]. Obviously this is multiple levels of OR. ;^)Minh Nguyễn💬02:19, 20 March 2023 (UTC)[reply]
Any coordinate provided in KML, or Lat-long, or UTM by zone, can be easily verified by either someone who knows Google Earth, or qGIS, or AutoCAD, or MicroStation, or ArcGIS. The idea that Google Earth can be anything other than slightly outdated is laughable to me: My entire career is built off the assumption that Google, Bing, ESRI, etc. are reliable indicators of the position of something on the globe to an accuracy of 1m. It's not 2004, the GPS system of Earth is far more capable than any secondary source that exists. - Floydianτ¢20:17, 20 March 2023 (UTC)[reply]
I think you need to reckon with the fact that the vast, vast majority of Wikipedians and Wikipedia users will not have ever heard of several of the programs you are citing. Let alone being able to verify anything using them. --Licks-rocks (talk) 20:37, 20 March 2023 (UTC)[reply]
And that's fine. The vast majority of Wikipedians can't verify a textbook from Estonia, but it's an acceptable source. The vast majority of Wikipedians can't perform calculus or calculate the Einstein field equations, but that's ok. They're verifiable claims, whether or not you are capable of that verification. - Floydianτ¢20:46, 20 March 2023 (UTC)[reply]
KML files are shapefiles, they do not contain any information about what was or was not visible on those coördinates at any point in time. If Einstein posted his field equations to Wikipedia himself, that'd be OR. And let me put this simply: I am not going to waste valuable hours of my free time trying to reconstruct whatever hair brained map another Wikipedian cooked up on google earth to prove that X or Y object did or did not have Y property at Z date. PS: Don't bludgeon this discussion please, you've had a new comment up every consecutive time I refreshed the page for quite a while now. --Licks-rocks (talk) 20:59, 20 March 2023 (UTC)[reply]
Yes, maps and photos can be subjective. The difference is, where disagreements on location arise, the aerial/satellite photos published in Google Earth are usually right, and the numbers from which Google's maps are made are mistaken. The placement of manhole covers and other long-lived infrastructure items allows great precision in urban locations, the main exception being in slant views. So, I use slant views to understand the environment generally and to identify objects, and straight-down views (where available) for precise locations of specific objects. - Jim.hendersonτ¢23:12, 19 March 2023 (UTC)[reply]
I agree in the real world, but Wikipedia is a bunch of amateurs. Is that tree-line a dike next to that swamp or is it a railroad ROW next to the old stockyards? And you can't fact-check.
I was always an "urban explorer" so I usually use the slant view and sometimes field-check. I've had pretty good luck with Google, it's an actual picture, of course it's accurate, but I have it a year or more behind in Forest View right now. I don't know how to date it. The map itself is only barely good enough to aim the satellite at a landmark and check street names. Sammy D III (talk) 02:05, 20 March 2023 (UTC)[reply]
I'm a utility engineer, and knowing where sewers, pipes and telecom conduits are in the real world is my job description. I've worked for several major companies that are the top of their field for this type of work. Google Earth is second only to the local municipalities GIS, and a physical survey of points. Both satellite and streetview indicate the date of the imagery, and I take that into account in my assessments (for example, if the streetview is from 2013, I send a team of surveyors out). You can argue that the placemarks/labels are wrong, that areas outside of cities in North America/Europe are displaced, etc. But you will never be able to convince any competent engineer that Google Earth is off the mark by anything more significant than what is required for construction. Wikipedia does not need that level of accuracy for anything in the built environment. - Floydianτ¢20:28, 20 March 2023 (UTC)[reply]
Hello to everyone involved on all sides of the debate. I have not read any of the discussion, but I noticed it in my watchlist after an editing dispute on a related page, and I think it is both needed and fair to give notice to all here that there is another related discussion now taking place at Wikipedia_talk:These_are_not_original_research#Old_section which is need of editor input for the purpose of deciding consensus. Your comments are requested. Thank you. Huggums537 (talk) 21:12, 20 March 2023 (UTC)[reply]
.png)