The tool is useful, but let down by Wikipedia images not being linked to the relevant Wikidata ID. I selected just the red circles (no photo). About half of these already have photos in the Wikipedia article that isn't linked to the Wikidata ID. Adding the image attribute to the Wikidata fixes it and the red dot disappears about 60 seconds later.
So before you head out with your camera, perhaps do some linking first. What's left surprises me with articles with no photo, which I would have thought someone would have done by now. So I'll be the first!
Hijacking the top comment to make it more clear since lots of comments here seem confused:
This map mainly shows Wikidata items, in addition to some "orphan" images on Wikimedia Commons (a free media repo from Wikimedia foundation, mainly for used on Wikipedia but not exclusively), both are not technically the same as "Wikipedia" articles.
- Red dot: wikidata item without photo
- Green dot: wikidata item with photo
- Blue dot: Commons images without a wikidata link (keep in mind not all images need to have a link; some are just too trivial to have any wikidata item.)
> not all images need to have a link; some are just too trivial to have any wikidata item.
Only the most representative image(s) should be linked from the Wikidata item. More precise properties also exist, e.g. you can separately link an interior image of a building or a tombstone image of a historical person.
OTOH, all images can link to all the Wikidata items they depict. This is using Structured Data on Commons which is basically an extension dataset linking images to what they depict: https://commons.wikimedia.org/wiki/Commons:Structured_data
Wikipedia only considers CC0, CC-BY and CC-BY-SA licenses free enough for the image to be included without any issues. With these licenses, you could also add images to Wikipedia that are not your own, which would greatly speed up coverage.
I'm specifically mentioning this because some CC licenses seem permissive, but are not permissive enough for Wikipedia. A key example of this is iNaturalist, which contains millions of photos of all kinds of species that would be useful to add to Wikipedia articles.
Their default license is CC-BY-NC. Most photographers on that platform just leave it to that default unaware or not interested in licensing in the first place. Some might be aware of this default license and think they're being permissive and generous, making possible exactly sharing across services like Wikipedia.
Unfortunately, no. It's a painful mismatch as many of those photographers would likely be perfectly fine with the image being used in that context, but the Wikipedia license requirements prevent this.
Similarly, even a license containing an attribution requirement (which is morally perfectly reasonable) is problematic. You find a photo from user "anon2522" on platform X that last posted 8 years ago. The attribution information is really soft and fragile, it doesn't really link to an actual person and the little information you have may disappear at any time.
Bottom line, if you're truly creating an image "for the good of the world", CC0 is highly recommended. You can still attribute yourself in EXIF or as a small stamp on the photo itself.
The clarification of "BY" <==> "Attribution" is highlighted reasonably well at https://creativecommons.org.au/learn/licences/, which also provides a comparative summary of the other CC licenses.
> You find a photo from user "anon2522" on platform X that last posted 8 years ago. The attribution information is really soft and fragile, it doesn't really link to an actual person and the little information you have may disappear at any time.
Is this really a problem? We've found stone tablets with authors' names on them from thousands of years ago; they're attributed, even if we know nothing else about their authors. CC-BY doesn't require being able to track down the person to be useful.
Yes, in general you can conclude that attribution is not a hard science, you attribute the best you can given available information. But even then there's painful gotchas as described in the article above.
The real problem I was aiming at is that it doesn't scale in the context of mass improvement of Wikipedia articles.
Say I'm an enthusiast regarding species of fly found in NorthEast America. There's thousands of species and we have thousands of Wikipedia stub articles lacking an image.
Next, I found a source containing said images and want to use them. One would now have to go through every single source image and do attribution to the best of our ability.
Is there a title? Mandatory to use. No title? Make up your own. Is a license included? If so, 3.0 or 4.0? Is the author mentioned? If yes, does the author want to be mentioned by their real name? Has the author included any other custom requests? You need to honor them. And so on.
It's incredible laboursome work if you take on the task of doing it properly. As a result, many may not bother to improve Wikipedia images on such scale.
Now imagine those images being CC0. You can skip every single step and be a 100 times more productive.
Blunt statement, but the most useful images licenses are "copyrighted" and "CC0". Any license "in between" is messy, not as permissive as you may think, and in many use cases (such as social media) their license terms are fully ignored anyway.
As someone who uses a lot of photographs in presentations, I try to attribute as best I can but it can't really be automated, a lot of information is frequently missing, and the attribution isn't directly connected to the photograph. So it mostly counts as a good faith effort but probably doesn't really serve the intended purpose.
The CC non-commercial variant is also very problematic. It's easy to understand why people use it. But it means that if you're being cautious, you mostly don't want to use an NC photo because most "interesting" uses aren't unambiguously non-commercial.
Why neither this nor tool in OP work with manually entering location is beyond me. My browser doesn't know my location. It doesn't help that it doesn't work even when I allow location in Firefox. But both would be solved if I could just select the location myself, as in any other application.
The tools use the browser location as it's easy to access in a few lines of code, and the users can select the location themselves in multiple ways:
* In the browser console
* In the browser settings
* In a browser extension
Meanwhile, addresses are hard in a global context. Implementing address search can be a lot of work because browsers don't provide one so you need either a (global) address database or an API endpoint to access one. Also, people expect a search which is both fuzzy and with autocompletion, which works best by prioritising results based on the browser location. Finally, you need to implement a user interface for the search field.
But it's extremely little additional work, though.
The application is already able to find articles around a given address. The only thing that would be changed, is that you enter the address by hand instead of letting the browser locate you. This can be easily done in a day
And it would be very easy to find out how many people start the page load, and don't complete. This will be close to the number of people who cannot or will not allow their browser to divulge that information.
Both TFA and your wikipedia special page rely on location access without any fallback for specifying a location if you don't want to give them your real location (or at least not to the precision that GPS allows) or if your are on a desktop where no location info is available. Disappointing.
Yes. In my case, it defaulted to a URL with coordinates in England. Tried editing the URL with localish coords, pressed enter ... nothing. Plenty of space in the toolbar to implement that. (Not the only map service that makes it hard to enter coords.) Also, don't see much need for more than 4 decimal-places after the Lat/Long values.
Seems like the range is fixed at 10km which gives me a total of 5 articles, none of them very interesting - and besides, I'm quite familiar with my area in that range.
Would be nice to be able to expand it so that I could discover interesting things a little further out but still in easy driving distance.
(Probably possible to do a sparql query for this at https://query.wikidata.org but seems like it would be complicated, at least given my relatively novice level of sparql skills)
What is the difference between wikidata and Wikipedia? Most of the red circles near me have a wikidata page, with no data, and no Wikipedia page. Plus to be honest 100% of them are business so I’m not terribly inclined to do their seo for them.
Wikidata is very inclusive. If there is a reliable public record of a thing existing (at the extreme end of the scale would be a park bench, tree, light pole) then it could be included. Wikipedia is quite selective with content included, generally requiring multiple public records documenting the thing, and for the thing to be something that the Wikipedia bureaucracy deem to be important enough to justify an article.
Hence Wikidata could have an item for every memorial park bench listed at [1] but Wikipedia just has a single page [1] describing what a typical memorial park bench is, and providing some examples either in list format or as a gallery of photos from Wikimedia Commons. Wikidata items for the memorial park benches could have properties for coordinates, official name, person or thing commemorated, date of construction, area the bench resides (e.g. park), length, height, width, construction material, identifiers assigned by heritage or government bodies, artist who designed it, benefactors who funded it, links to significant events related to the bench (e.g. used in a movie), etc.
Wiki data is meant to be some universal facts source that things like voice assistants to tap in to. Answering things like “how tall is Obama”. Wikipedia has much stricter restrictions on “notability” so most businesses won’t quality for a Wikipedia page but would be accepted on wiki data.
Would it be easy to make this tool work for people who don't want to mess around with browser settings? Or would it be easier for me to learn how to tell the browser where I am?
Yes, it definitely needs a fallback for browsers that have no idea where they are. That would also make "Find Wikipedia photos needed near where you are going to be" possible.
Edit: The search icon more or less works, but just showing a navigable map would make much more sense.
This is so awesome, I just discovered the large rock feature near the highway about 200 meters from my house is called "Chipper's Leap" where in 1832 John Chipper jumped from the rock while trying to escape an attack by a party of Noongars.
If I look in my region it is absolutely filled with requests of pictures of every single street and canal making it very hard to find something that's actually worth taking a picture of.
There are notable buildings here, some statues. Not sure why you want to have a picture of a random street where some people live, I mean the privacy issues alone. At that point use streetview.
It would see some analysis of overlapping red and green circles (or image titles) would be beneficial. Near me there are quite a few existing green circles but also a red circle asking for the same picture.
It should be noted that toolforge tools are run by volunteers, and are subject to a different privacy policy than the main wikimedia sites (that said, they are not allowed to do shady things).
You should only need an account on wikipedia not toolforge. Blame all the people who keep uploading pirated movies.
"Vetted" is a bit of an exageration (its pretty easy to get an account. Sure technically there is a process, but its an easy one).
Basically you can't openly be evil, and if someone discovers that somebody is being evil they get kicked out. Also i think there is a proxy so you can't easily collect ip addresses.
So its a step above some random website, but ultimately its closer to a third party site then it is to being real wikipedia.
Given you are all on hackernews which is basically a link aggregator for random websites, there is really no reason for this to give anyone pause.
All the other reasons like spam notwithstanding, without at least some persistent user ID, how would image licensing work? You'd only be able to have works that can be shared and modified freely without attribution.
Wikimedia is surprisingly lenient on accounts, you don't even need to give an email to get a username.
Yeah, I found the login requirement very confusing, and I definitely wasn't about to grant high-volume editing permission to an app I'd never heard of (WikiShootMe is the URL, but it's not mentioned on the page or in the HN article title, and HN clips off the subdomain).
I needed to go hunt down the documentation (available under the hamburger menu), which makes it clearer (eventually) that authorizing is only required if you want to upload pictures through the WikiShootMe app - so you could skip authorizing and upload through Wikidata/Wikimedia Commons instead.
I tried using it several times in recent years, and whenever trying to upload files, I had lots of issues. Filed some bugs, never got them resolved, and ended up giving up after several of my photos and descriptions were lost (or never properly uploaded).
I really want the Commons Mobile app to work, and I don't know why it always failed with my Pixel phone, but it was a really frustrating experience.
Interesting. Though a lot of the pages near me only exist on non-English Wikipedia (which is interesting since I'm in the United States). I'm not sure how many people where English is not the primary language will be searching for a tiny dam or bridge in a mid-sized city in New England. I'd want to make sure my photo has impact before I spend time driving to the location and taking one.
Do you have any examples of how photos with location are currently being used? I'm all for Wikipedia and sites having this kind of data to build on. But who already built on it?
I was thinking more small scale. Someone wants to make an educational youtube video about some obscure landmark and needs a picture but doesn't want to get copyright struck.
Same, I'm in Vietnam but most of the pictures seem to link to Spanish Wikipedia which, apparently, has an exhaustive range of articles on beaches around the world.
It could be that other languages are restrictive than English Wikipedia which is notoriously strict on what articles they allow.
There's a school next door, and the school (the building) has a picture, but the school (the administrative unit) does not have one, even though the school (the administrative unit) is listed as residing in the school (the building). So this tool shows that the school (the administrative unit) is missing a picture.
This is really neat, just as a way to visualize things near me with Wikipedia articles (even if many of them are buildings that don't exist any more). It's also interesting to see where the articles exist. For example a random building near me that was built in 1910 and burned down in 1913 has articles about it in English (not surprising since it was in Los Angeles) and Croatian (somewhat surprising)
I was able to search just fine without logging in to anything. Denied location access, manually entered location, map popped up with blue, green, and red bubbles.
Fun discovery: some of my own pictures are up there around my house! I've documented hundreds of species in my garden using iNaturalist, with all photos CC0 in the hopes that they'll be useful to someone. Neat to see that some at least are getting found and used :)
I'm confused about wikidata vs wikipedia - I'm surprised by the large number of dots near me, I thought every dot was related to a wikipedia article, then I clicked on the dots and realized they were wikidata, not wikipedia
I have a personal opinion that might not be popular
Wikipedia is actual encyclopedia and useful
Wikidata is a weird idea to create a weird “semantic web” - a hierarchy of everything that ever existed or will exist - and it’s about as useful and easy to understand as previous experiments with semantic web.
Wikimedia seems to double down on this instead of letting it die; I really don’t understand the point, I don’t think that it’s useful and it’s really hard to edit, as a editor.
Interesting thing is - it’s CC0 instead of creative commons for some reason. Their argument is that you cannot copyright a fact. I donno.
Weird take when wikidata has 117% year-over-year page view growth.
The audience is quite different, and it may very well not be useful to you, and that is ok, but wikidata clearly has an audience and seems to be on an upswing. Its certainly doing better than pretty much all the other sister projects.
As an aside, politically wikidata is really WM-DE's baby, so its really more them doubling down and WMF along for the ride, but of course its all very intertwined.
It’s definitely not the least useful of the wiki projects - that award definitely goes to WikiNews - but at least WikiNews is not intertwined with everything on wikipedia.
Thanks for doing this! I am going to contribute some photos soon, I've added a todo for this weekend. Plenty of red circles near me, and it will encourage me to see some new places in my area.
I get a blank screen with no way (that I can discover) of entering my location. I tried clicking the spyglass icon and entering my city. But just blank results on that as well.
It sort of breaks. If you search for something the map will load around Denmark, at which point you can zoom out and manually navigate to any location.
Rather it's "find wikidata photos needed near you". None of the stuff near me is remotely significant enough to have its own article. I don't understand why anyone at Wikipedia is interested in storing photos of such things as unused twenty foot fishing ponds and tiny back yard creeks.
Please don't complain about tangential annoyances—things like article or website formats, name collisions, or back-button breakage. They're too common to be interesting.
There really ought to be a map legend on the page itself.
EDIT: After reading the documentation page [0] linked in another comment, I see that there is legend in the "layers" popup on the upper-right. It did not at all occur to me to look in this location; I would have guessed that this icon let me change the underlying map tiles (e.g. transit, traffic, topography, etc).
I stopped to contribute to Wikipedia after they banned edit access from VPN services, even if you are logged in, your account is old, and has no spam history in the past.
And how much will they pay me for this work? Considering they have tons of money to spend on things unrelated to Wikipedia, yet each year asking for more in their puppy eyes boss scam.
I consider Wikipedia to be one of the crowning achievements of the Internet, this isn't my worry at all (despite being annoyed by their funding campaigns, too)
So before you head out with your camera, perhaps do some linking first. What's left surprises me with articles with no photo, which I would have thought someone would have done by now. So I'll be the first!