2009/10/7 Matthew Somerville <[email protected]>: > > One would assume the Ordnance Survey have taken legal advice on their > OpenSpace, which appears to state that point based derivation from an OS > map would belong to the OS: > http://openspace.ordnancesurvey.co.uk/openspace/faq.html sections 2.1 > and 2.2; and
2.1 (which is the relevant one) is suitably weasely. "When you use OS OpenSpace to geocode data by adding locations or attributes to it that have been directly accessed from and/or made available by Ordnance Survey mapping data, then the resulting data is 'derived data', because it is derived from Ordnance Survey data." If you take a mapped feature and use OS OpenSpace to create a (location, attributes) tuple, you are doing two things: (i) if you do this to all such features, or very many features, you are doing something akin to tracing a copy of the map and adding additional information, that would be a derived work just as much as if you used a pantograph or tracing paper; (ii) the OS may have a database right in the features that they present, if you extract (a substantial amount) of them then you may infringe any database right they have in them. Looking at a map of my area which does not show my house (which quite a few online maps don't at some resolutions) and clicking on where I know my house is, does none of these things, because I am not copying any element of the map etc. 2.1 could be read either way (and that's how I would draft it if I were the OS) but I don't think it is in any sense conclusive, even if they were right of course. > http://openspace.ordnancesurvey.co.uk/openspace/developeragreement.html > has their definition of Derived Data which includes "any Data created by > You or an End User identifying the location or other attribute of any > new feature directly using Our mapping". A separate point to do with licensing not intellectual property. It may be (depending on licence) that what we are talking about would be a breach of someone's contract/licence agreement somewhere. That's a separate (and more easily circumvented) problem. Its usual to claim as much as you can in such a licence. > > I did misunderstand your previous email (obviously if you used Google > maps search to look up a postcode, then clicked on the map, I imagine > that would be a worse position), and I do agree with what you say (my > postbox locating thing revolves around a presumably similar issue). Absolutely - using (say) code-point to find the location of a postcode and then getting a user to confirm it would be an extremely clear infringement (because you are just copying their database, with a human check stuck in there). It may be that there is no way of setting up such a thing so that it would avoid these problems, but I think that it probably is. Of course using OSM as the underlying map would avoid that problem. But all this would be disastrous because the data wouldn't be accurate, complete or always up to date (yes, I know the PAF isn't either, its not just that it can be out of date it contains errors and duplicates too). Why crowd-duplicate what the Post Office are already doing at our expense. -- Francis Davey _______________________________________________ Mailing list [email protected] Archive, settings, or unsubscribe: https://secure.mysociety.org/admin/lists/mailman/listinfo/developers-public
