What the examples don’t cover is the retrieval of ordinary facts from such images, for example the material from which some hut is built, but even that was considered non-permissible for OSM
“Was considered” by who? (thanks for sharing that conversation, so anybody can understand this point of view. Especially if it’s not distant from the problem described in the top-post of this thread, which is about saving a link)
I am not a lawyer, but even if your claims would be all true then it does not ban all commercial use
(yes, some categories of commercial use would be quite doomed in practice, more if your claims are accurate, but some income-generating activities compatible with license are 100% fine)
this claim is simply wring, under any reasonable interpretation
can you edit it to clarify it? At least
“OSM does not require a share alike” to “Images linked in OSM data does not require a share alike”
and
“Anyone using OSM for commercial purposes would violate the cc-by-sa license.” to “Some uses for commercial purposes would violate the cc-by-sa license.”
?
I have made a similar comment to Minh in the past.
In the real world there no such thing as those absolute truth statements you made. What is true is that OSM takes a very cautious approach to utilising potentially protected material from third parties, avoids grey areas as far as possible and will ask for permission when in doubt. The issues with tracing from aerial imagery have been discussed many times, and Minhs position is not that of the OSMF LWG that is, for the most part, actual lawyers (which should be speaking for itself IMHO).
I hope I didn’t quote you wrong, that would have been unintended.
For me that follows from Minh’s explanation about the wall in hot air ballon example. I might be wrong.
Will do shortly, that was maybe shortened too much. Commercial use is certainly not forbidden by cc-by-sa 4.0 but may be doomed as you say.
I’m aware that opinions and interpretations about close to everything in OSM, let’s say, vary quite bit. So I take advice from people with lot’s of experience more than mine as advice. It would be helpful if OSMF LWG would provide statements about the dos and don’ts in concise form, that is, look-up table like. If this is to be found somewhere, someone point it out, please. That this thread exists, however, suggests there is no such table.
Sometimes it appears to me that these copyright discussions are an expression of the desire for the laws to be different than they seem to be and to reinterpret them in workaround fashion. @SimonPoole Can you tell what information I am legally allowed to read from a cc-by-sa 4.0 image (of a building with a hot air ballon 
) and enter it into OSM and what not?
On the tag level. Does OSM regulation allow to retrieve:
housenumber
levels
material
roof shape
surface of the road in front of it
etc.
from said image on Wikimedia?
I hope the OSMF’s lawyers say yes, you are allowed to do so but on any previous occasions when asking such questions the answer was no. Is there a quotable statement on this somewhere, then we could maybe close this thread.
Copyright protection should of course not depend on the bank statement of the holder.
only some specific cases (for example, good luck selling digital copies of CC-BY-SA 4.0 licensed images)
some will be perfectly fine, I personally used some CC-BY-SA 4.0 materials in commercial activity while fulfilling their license (more than once)
Nobody, including Minh before he went off on a tangent about aerial imagery, has claimed that from a copyright perspective there is any special permission required to determine factual information from an image that you have access to on open licence terms.
But that has no bearing on any of the other questions, if you display the image, create a derivative etc, you will still need to follow the licence terms and naturally you might be subject to regulation outside of IP law, say for example data privacy laws, or by private contracts (see the google SV example) that will stop you from using that information for mapping in OSM.
I apologize for the confusion that this paragraph caused. My intention was to emphasize that not all copying is copyright violation. I realize now that this tangent about semantics was unhelpful. The more important points I tried to make before and after this paragraph are:
This isn’t legal advice, it’s a best practice for members of the community who don’t want to get dragged into messy discussions. If you’re doing something especially risky, like a mass import or a software feature for automating the mapping process, you may need to get a second opinion.
As far as the original question, I thought that had already been settled. It should be pretty clear that linking to copyrighted materials is far from unusual.
If you get mixed messages from me and others here but you need a more definite answer, you could contact the LWG for guidance.
If I were to use an image with cc-by-sa 4.0 in my book, I would have to provide cc-by-sa 4.0 with the reproduction of the image in my book.
Yes, the image (and any derivates thereof) must be distributed with the same license. CC BY-SA is not “viral”, it will not automagically need the whole book to be licensed CC BY-SA.
If I use an aerial image and sketch the paths on it and publish that sketch, I “build upon the material” and have to publish it under cc-by-sa 4.0.
No. If you use an image to determine the name of a specific store, would you need to attach a CC BY-SA license to that information? Obviously not. Neither a store’s name, nor the paths visible in a photo are a individual creative expression that is subject to copyright.
The beliefs surrounding the limitations of basing maps on (especially aerial/satellite) imagery in the OSM community are more dogma than based in actual legislation.
The beliefs surrounding the limitations of basing maps on (especially aerial/satellite) imagery in the OSM community are more dogma than based in actual legislation.
That is not quite correct. I’ve been asking people working in GIS offices (Germany) that produce aerial images as well as maps and who know about the respective licensing situation. According to them, you can not (in Germany) just use whatever imagery is online to create maps from the pictures but need a license (unless stated otherwise). In some federal states the use of specific aerial imagery has been permitted in general or to OSM specifically and some of this required lengthy negotiations between OSM and the state. In some others, it is still not possible because their conditions for usage of the images require naming the source for each use case, which OSM can’t do.
Copyright holder may refuse to give a license even if it isn’t required by law, and fact is fact: permissions for tracing of aerial images is rather measure of the precaution than actual requirement (licenses for images on the wiki are required, but who cares!).
That’s not (the fact that they make the imagery available under licences that may or may not allow reuse by OSM) the point in question, it’s if a) aerial imagery is even covered either by creative copyright or sui generis database rights, and b) if tracing from such imagery is use that would violate any exclusive rights granted by the law (see transformative use in German Urheberrecht).
There is as far as I know no case law that is directly applicable, so it’s all a bit up in the air. That’s why we take the safe route and abide by the terms of their licenses not because we think your average employee of a government GIS agency knows what they are talking about on legal matters.
Note that this is very jurisdiction dependent and for example the answer in a sweat of the brow doctrine country (UK) could very well be different, not to mention countries that have special legislation that provides copyright-like protection to government geodata (CH).
I would be very interested in whether their guidelines are actually grounded in copyright legislation or rather in agency policies. I do understand that OSM tends to err at the side of caution and that it prohibits use of factual information that can be derived from imagery depending on the source. I can (to some limited extent) also kind-of follow the opinion to only allow to use imagery where there exists an explicit permission to use said imagery to derive factual data from it. Nonetheless, this is more of a belief system than actually being grounded in copyright legislation existing in the real world.
I find no issue with OSM following an approach that wants to err on the side of caution. I do, however, object strongly and vehemently to OSM being a source of disinformation and sectarian beliefs as they refer to copyright. Do roll your own limits, but if you disseminate them as if they were gospel, you are false prophets and support the wrong side of history.
…in particular, the said “belief system” is mostly grounded in a (mis)interpretation of the concept of derivative works, whereby an act of seeing a CC-licensed photo and then translating the objects encountered there into database entries constitutes creating a “derivative work”, which in turn requires attribution of the original authors.
In the real world, I think these questions would be answered with a risk-based analysis, while in OSM, mappers want a definitive list of absolute dos and don’ts. That list is naturally going to err on the side of caution, due to a long tradition of map publishers laying out traps and terms of service that most of us can’t comprehend. Is the clean room mindset strictly necessary? 
Of the five or six cases that the original post brought up, four were about linking, which has always been considered acceptable in OSM. The case that led to the rest of this discussion was:
The example sounds innocuous enough. It’s fine, right? But be careful with extending this line of reasoning:
Suppose the images on the website aren’t under a free license and they’re actually scans of static maps rather than photos. In that case, it’s possible to glean facts from the images without any problem, and it’s also possible to violate the maps’ copyrights, regardless of whether text is involved.
Fortunately, that wasn’t the case in the original post. But having already answered the original question, we started discussing the broader implications:
So, setting aside our belief system, should we document on the wiki that it’s actually acceptable to use Google Maps’ satellite view and Google Street View to map houses and their addresses? Or maybe the answer is still no, but not quite because of copyright?
Two things can be true at the same time. Photos from Wikimedia Commons can be a great source for mappers, and photos from Google Maps can be not a great source for mappers. Do we agree on that much at least?