3 ms·
The issue here is one of rights. Who owns the rights of the generated data? Especially if it was the "crowdsourced" product of n people. Does the data make any
by ris 8y ago
The issue here is one of rights. Who owns the rights of the generated data? Especially if it was the "crowdsourced" product of n people. Does the data make any sense outside the context of Umbra's software? Is it exportable? Or do Umbra retain the rights to take their ball away at any time?
These are all questions anyone should be asking before putting any sort of time & effort into "contributing" to some sort of "crowdsourced" project. Or even putting a significant amount of work into a cloud-based tool.
Mapillary's answer was a compromise - contributed image data is published under a creative commons license, but Mapillary retain copyright over the metadata (most importantly positioning information). This is a bit of a fudge which would prevent users reclaiming the entire dataset and taking it elsewhere, but it seems to work for the community ... mostly.
- candiodari 8y agoGenerally this is considered a compilation. And the compiler owns the rights to a compilation. The biggest example would be the yellow pages (let's assume you remember those). The company putting it out certain owns copyright over the book, it's exact contents, even if it owns none of the individual details inside of it.
- daveguy 8y agoI'm not sure why this is being downvoted. Is the statement just not correct? (seems right) Or does it vary significantly throughout the world? Or is it down voted because people don't like the idea of their collected information being owned by the compiler? There has to be a lot of grey in this space. Can someone clarify?