This blog post originally was originally written for FutureEverything as part of their Open Data Cities programme.
I’m no expert but I really need to be — Licensing
Licensing is a subject that comes up a lot with Open Data. The licence is a key component of the dataset. It defines the use and liability and it shapes how or what innovation will come from data release.
As mentioned in the title I am no expert in this area and I would appreciate any correction or amendments to my understanding.
Traditionally public data has been closed so that the only way you could get access to data to build products was by buying a licence to use. In many cases these licences were expensive and restrictive. The to mitigate this cost often, te licence would also have some level of service agreement built in. You paid for the licence for the data and the data provider would provide you with a level of continuity and support. This helps to limit risk and encourage investment into a product.
The closed ‘paid licence’ system generally has a high barrier to entry ‘price of licence’ limiting the amount of innovative products developed. If innovation ecosystems are ideas that live with most failing. The price of failure being too high could have a chilling effect on the whole system.
One of the first licenses used for the release of Open Data was Creative Commons CC-BY-SA. This licence allowed people to create services and products off the back of the data as long as they attribute where the data came from and share back any data that was created off the back off the originally released dataset (value-added data). The original Creative Commons licenses were devised as an answer to restrictive copyright laws relating to ‘works’ — articles, text, images, music etc., as these were deemed increasingly anachronistic in the digital age. It is up for discussion if data can be deemed as a ‘work’ in the context of this licence.
The Open Database Licence (ODbL) developed by Open Data Commons, was created to address the doubt that data could be seen as a ‘work’. It carries the same attribution and share alike clauses and is used by many datastores including the newly opened Paris Datastore.
Anyone can develop products and services that use datasets with these licences but intellectual property doesn’t extend to the value-added datasets created in the process of developing these products. Releasing value-added datasets back to the community allows further innovative products to be released off the back of these datasets, so potentially the pace of innovation could be increased — It is analogous to the ‘standing on the shoulders of giants’ idea.
By imposing further use of value-added data by other organisations might chill the development of products that create value-added data.
With the above licences there is generally no liability or guarantee of service from data providers. This creates a greater risk scenario. If you were investing in product development this potentially is a source of concern and may be an inhibiting factor
In the UK we have the recently released Open Government Data Licence. That was developed specifically for government data. It borrows from some aspects of the CC-BY-SA licence and ODbL. Unlike the those licences there is no need to share back value-added data.
Would this have any impact on products and services that are developed from Open Data? Again in the licence there is no liability or guarantee of service from the data provider but the developing organisation gets to keep all the rights on the products and services they develop — including value-added datasets.
The advantage of this could be that by allowing people to keep hold of the rights to the products that they develop might be mitigate against the exposed risk posed by the lack of liability and guarantee. The main disadvantage could be that the pace of innovation could be curtailed due to people having to replicate process and value-added datasets.