there was a project to import Geograph images into Wikimedia Commons, so that’s arguably even better (but you can’t find every image there):
That’s amazing, thank you so much!
there was a project to import Geograph images into Wikimedia Commons, so that’s arguably even better (but you can’t find every image there):
That’s amazing, thank you so much!
You’re welcome, happy mapping!
I’m not a lawyer, but I suspect that regardless of the licence there wouldn’t be a legal issue just linking to an image. See here and read down to the bit about linking.
However, anyone consuming image keys from OSM might then have the problem, so it does make sense to ask the question.
I thought that there had been specific discussions around Geograph in the past though - might be worth a search of the talk-gb mailing list.
I thought that there had been specific discussions around Geograph in the past though - might be worth a search of the talk-gb mailing list.
Is there a better way to search for it than just CTRL+F’ing the talk-gb archive?
This search gets a bunch of posts. I can’t guarantee you’ll find the answer to the question there, but it’s worth a look.
The ODbL is a database license, not a content license, thus it is not meaningful to say it is or isn’t ODbL compliant. There are no legal reasons why you can’t put a URL which contains a restrictively licensed image into an image
tag. Putting the image itself in would be an issue, but it would be very hard to put an image into OSM with the OSM data model.
Woudn’t any OSM-based app which makes use of the URL to show the image, be subject to the license requirements and restrictions? If so, wouldn’t that compromise the free and open use of OSM?
It would be different if the URL brings the user to a different website where the image can be found, subject to the license rules of the website.
main question is answered in Is it allowed to add a link to a non-ODbL compliant picture to an OSM object? - #11 by pnorman
according to my knowledge, note that I am not a lawyer, yes
(that is why wikimedia_commons=File:Gleann_Miabhaig_-_geograph.org.uk_-_5143695.jpg
is superior, with image
links you approximately cannot do this due to lack of machine-readable license data)
just linking is fine (some countries introduced laws that ban linking in some cases, though AFAIK even that would not apply here)
no
license of OSM database applies to OSM database, not to everything mapped or linked
Manually, there are some ~2.6k https://taginfo.openstreetmap.org/compare/image:license/image:license:wikidata
~37.7% of image=
vals are related to wikimedia_commons=
, mapillary=
, flickr=
, kartaview=
(including =*openstreetview*
) , bubbleid=
, or panoramax=
. They can be ignored.
note that it is tiny compared to image
use in general and work of a single mapper (and having this images on Wikimedia Commons would be in general much better idea…)
No. For example, you can see the images by clicking on a link on OSM.org. If embedded directly there are more issues, and if the use is one that requires permission from the copyright holder, that would be an issue.
Presenting images from random third-party sources is a terrible idea anyways, as you’ve now opened the attack surface of your program up enormously.
If you embed a website from a website
tag in your app there are similar issues. Being told where to find external content does not change the license OSM is under. If it did, it would be an issue for all the Wikipedia links.
I would not recommend using an image tag in any case, as it’s poorly designed. If the goal is to provide source information other mappers can look at, it should go on the changeset tags like source does. If designed to be data about the object, it is useless. Two mappers might see the object and decide on different images.
Mapping always involves judgement, but this is hopelessly vague. Do we put multiple useful images in the tag separated by ;
? How do we judge what a useful image is? If there’s more than a few, the URLs combined will be too long for a tag.
Something like wikidata is usable because most of the time there’s one or zero correct values, and only some cases is there more than one wikidata ID for a single OSM object. Two mappers will generally arrive at compatible results.
Some general remarks about Geograph (the subject of the OP):
This is generally a good practice. If you have a photo of an object the coordinates of the photo may be of where the camera was, or where the object was. If it’s the camera, it’s difficult to work out the coordinates of a point in the image without referring to imagery.
I agree with you technical criticism but think that it should be reworked instead of never used.
Here I disagree. The reason why I was asking this question is so that I can add images to remote location on/near hiking (long distance) paths like the West Highland Way or the Cape Wrath Trail. The images can (and will!) help a hiker especially during the route planning phase to spot good places to set up camp or whether a hut is a viable sleeping possibility (not all are on the MBA website), etc.
The specific problems with the image
tag (or wikimedia_commons
, etc.) do not really arise here, as there is (most of the time) just nature and one POI and a single photo of it.
Should we expand it with links to the relevant discussions on here? Maybe a short (neutral) summary:
Damn, that’s cool, didn’t know Germany has this as well.
Does wikimedia_commons= have a significant advantage over image here=, in that it can point to a category? So another mapper with a more recent photo could point to a category including both your and their images. With image= they would have to decide if their image is “better” than yours, or if the location has changed enough to justify replacing your photo, or if a winter image is more useful than a summer image, or…
Yes, I’d say that’s an advantage, for things that are photographed enough to warrant a category. The existence of a category on Commons automatically qualifies the thing for an item on Wikidata, which can link to Commons and feature the most representative photo within the category. Some applications like OsmAnd already know how to discover photos via Wikidata. So ultimately wikimedia_commons=Category:*
is merely a placeholder for a future wikidata=*
tag, but a pretty useful placeholder for someone who doesn’t have the time to create a bunch of items.
Recently there was an extensive discussion about the relationship between image
and wikimedia_commons
that touched on some of the points made here:
I don’t see how this contradicts @pnorman . I might choose entirely different images in identical locations.
If photos are accurately geolocated and hosted on a site where that can be queried easily there is no need to add that information to OSM (see how cycle.travel does this). It also avoids the issue of ensuring an image tag is updated if something like a bothy is destroyed or taken out of use. If I walked the Cape Wrath ‘trail’ when rivers and streams are in spate then I may want good photos to choose a good place to ford them, but these places change, the underlying imagery where the river crosses the end of Sandwood Bay is quite different from what is mapped.
But adding a wikidata
or wikimedia_commons
category is surely of use?
or wikipedia
tag.
as long it matches and it is not linking ceb
botopedia or Wikidata entry generated for ceb
article - yes.
( Cebuano Wikipedia - Wikipedia - basically entirely bot generated, wild duplication between their wikidata entries and of human-edited Wikipedias)