The Task Force work has been divided into two broad areas – exemplars and enablers. The ShowUsABetterWay competition has taken forward the exemplars strand. On the enablers side, we have been working on a model for a different architecture for public service information.
Our enablers work looks at what is needed for a usable Access Layer according to this model. I have tried to illustrate this with some simple questions from a re-user perspective. These are:
DISCOVERY – can I find the data that I want?
LEGAL – am I allowed to use the data?
TECHNICAL – is the data in the right format?
COMMERCIAL – can I afford the data that I need?
Discussions at the workshop generated some additional questions that we also need to ask of the access layer.
INTELLIGIBILITY – can I easily interpret the data that I am accessing?
DEPENDENCIES – does this data depend on anything else that could affect my use of it?
This last point on dependencies is drawn from the free software world where software is bundled into packages that are managed by tools like apt for Linux. Each package is constructed so that it is aware of its dependencies on other software packages.
Package libraries have grown up for related software families like the Comprehensive Perl Archive Network, CPAN, and this model has been deliberately adopted for the Comprehensive Knowledge Archive Network, CKAN.
The discussion that is ongoing around the use of geographical information that can be regarded as ‘derived’ from Ordnance Survey data is a very good illustration of the need to have dependency information associated with any public dataset.
Data that is dependent on Ordnance Survey data may have very different re-use characteristics from data that has been constructed independently. These significant differences may also apply in respect of other dependencies, for example on Royal Mail Postcode Address File data.
Richard Allan, Task Force Chair