OpenStreetMap
There's one thing that came to me regarding the People's Map gorgeous (but unavailable to us) imagery. Basically, the principle is simple: you can't use it to map directly, but you definitely can use it to spot missing things. So basically what I'm doing is looking at the close-level imagery to spot missing streets or footpaths, and then entering OpenStreetBugs for people to map them (Yes I know, I should map them myself, but I live in Montreal, so that wouldn't be very practical :p).
Discussion
Comment from Richard on 27 April 2009 at 19:09
I believe there was a court case in Singapore fairly recently where this technique (though wrt maps, not imagery) was judged a copyright infringement.
Comment from Circeus on 27 April 2009 at 19:21
A quick search reveals more muddled facts (the case is Virtual Maps v. Singapore Land Authority, BTW) AFAICT, Virtual Maps claimed the SLA data was used solely as a normal roadmap when gathering data, but the court did not rule that THIS was infringement, but merely that howevr it happened, there had been infringement (notably noting the presenc eof SLA easter eggs in the VM data).
Besides, I thought we could, for example, use name plates visible in Google Street as data sources?
Comment from mapryan on 27 April 2009 at 20:28
All sounds like derived works to me, which I thought was a no-no.
Comment from LivingWithDragons on 27 April 2009 at 20:40
I don't think it's helpful to map somewhere that you are quite far from. It is quite possible someone has gone around there and checked out that things don't exist (aren't public or aren't what they look like).
To copy signs from Google Streetview would be a derived work. The thing is Google own the photos (but not the map data), if:
a) they allowed us to (It's verbally bee told to me that we could, but I think in writing is important)
b) we copied the names *without* using the overlaid road data which is owned by Teleatlas/Navteq/etc licensed to Google (how do we know where the road sign photo is, without comparing it to the named road layout of clicking the map to start Streetview? So perhaps we need Google to give us an api for accessing *their* data without the maps, but I don't know if they want to do work just to help us)
Comment from Circeus on 28 April 2009 at 04:29
Last I looked you couldn't be sued for noting facts represented in a photo (in fact, I distinctly remember my google exemple being distinctly mentioned somewhere as a way StreetView could be used). This would be like a T-shirt maker asserting copyright infringement because I noted that they sell T-shirts with the text "So and so" on them!
Remember that facts (i.e. the existence of footpaths) CANNOT be copyrighted, only their representation (i.e. the specific mapped representation).
Comment from djo0012 on 28 April 2009 at 04:41
just a note about street post in google street view... did you actually tried to read one? I just did a few km in Portland and I couldn't read any at any zoom, so maybe if it's not copyrighted it can help for thing like postbox footpath kinda I mapped a bit of my city and did it with picture to help me remember the thing I wanna map, but I forgot which direction I was looking while taking some pict, that's something were street view might be interesting, but it's not good for street name.
Comment from Circeus on 28 April 2009 at 04:44
I actually did a long while after street view came along. The images have ridiculous amounts of glare, and indeed all I could manage was to confirm a) that a street did not go as far as the map claimed and b) that it actually had a name (you could see, if not read, the name panel).
Comment from daveemtb on 28 April 2009 at 08:23
I thought OpenStreetBugs specifically states that you should not enter bugs based on licensed information?
It is best to avoid any grey areas on copyright IMHO. OSM is based on a conservative approach to this issue. There are many shortcuts that have been avoided so that we have data no-one can claim is infringing copyright, and to make sure the whole database stays pure. Look at the caution that was excercised before approving use of Yahoo photos: http://wiki.openstreetmap.org/wiki/Yahoo
Comment from daveemtb on 28 April 2009 at 08:25
Oh, and on Google Street View, I agree that taking facts from photos would probably be ok, except for the fact that the photos are tied to copyright map data. You are refering to the location based upon this map data. You can only follow roads on these maps.
Comment from daveemtb on 28 April 2009 at 08:48
For comparison, see the section on Wikimedia Commons Photos at: http://wiki.openstreetmap.org/index.php/OpenStreetPhoto
I agree with this point of view, and if Wikimedia Commons geotagged photos are tainted, then GSV certainly is!
Comment from Circeus on 28 April 2009 at 13:21
I take some issues with the OpenStreetPhoto analysis (where the hell is the ground for assuming any and all geocoded photos had their coordinates taken from Google maps?). Second, people keep forgetting that there is no "breaching of Google's copyright" given that (AFAIK) very little (until they added their mapmaker, only the usermade map layers were (c) google) of that data is actually owned by Google (i.e. virtually all their U.S. data is the TIGER dataset as licensed by Teleatlas). Now a breach of a PATENT over Google Street View is another issue entirely.
Comment from Harry Wood on 28 April 2009 at 14:37
You're arguing points which have been argued many times before. That's OK. Various people will debate the ins and outs of copyright law with you quite happily, and you can state your interpretations (interpretations which may or may not be correct / stand up in a court of law) but none of that really matters because...
That's not the way we build maps for OpenStreetMap. We do not use any copyrighted sources of information no matter how indirectly. We don't trace over them. We don't take measurements or pinpoint coordinates on them. We don't use them to find missing mapping progress. We don't even look at them whilst working on our maps.
Anyone who does these things is damaging the integrity of our open licensed data. Anyone who states publicly that they do these things, and that other might follow their example, is doing even more damage.
It's fine to have discussions about copyright law, and how it applies to our mapping techniques. This might even yield some new data sources or techniques that in the end we all agree we can use. Reading signs in google street view might end up being an example of that (although is this discussion still just based on a casual remark from Ed Parsons on twitter?) but for the time being this all just discussion. Please don't put into practice, or make statements about it being fine to do so.
Comment from Richard on 28 April 2009 at 15:46
Er, the OSP analysis is based on the fact that Wikimedia actually tells you to derive from Google Maps. That's pretty unambiguous.
Harry and Dave are right. If you want to prove what you think is legal about copyright, then fine. But do it somewhere else, not with OSM. OSM's policy is, and always has been, that we don't take the risk. It's unfair, to put it mildly, to put others' work at risk just because you're personally convinced of something.
On the specifics:
The argument that "facts CANNOT be copyrighted" is a wild over-simplification and if I hear one more North American recite it parrot-fashion I will not be responsible for my actions. In some jurisdictions, collections of facts can be, and are. That's what the EU database right is. Yes, it's a shit situation. We don't make the law but we have to live within it.
For Google Street View, my personal view is that it could be legal in some jurisdictions to take street names from it – but with two big provisos. The first is that Google's Ts&Cs don't prevent you (i.e. breach of contract), and they probably do. The second is that you certainly wouldn't be allowed to use it in conjunction with geocoded information; if you recognise the street from the photos, perhaps you could then take the street name, but not if you were using the geodata to identify which street it was that had this name. Even then, even despite both of those, I'm not entirely convinced any of this would be legal in a sweat-of-the-brow jurisdiction like the UK.
Comment from RichardB on 28 April 2009 at 16:38
Can I suggest also that the Getmapping imagery, as I recall, was created in or around 2000, so will be getting on for a decade out of date.
Apart from the usual things that are hard to survey on the ground, rivers, field boundaries, building outlines, landuse areas perhaps - it's not going to be much better than actually going out with a GPS and mapping properly. Some features on the imagery are probably no longer there. There will be other features on the ground now that aren't on the imagery.