Data Release Processes
From Open NZ Wiki
The State Services Commission has published a draft New Zealand Government Open Access and Licensing framework (NZGOAL). The framework recommends a Review and Release Process that State Services agencies can follow to manage policy, legal and practical considerations around releasing data.
The World Wide Web Consortium (W3C) has published a draft guideline for Publishing Open Government Data.
I have raised a couple of issues pertaining to the use of CC3 for licencing data, The following is the text of a comment I sent to NZGOAL describing these. Comments on this issue from others here would be useful
Brent Wood
A note describing what I see as severe limitations regarding the use of CC3 as an effective licence for open datasets in general and spatial datasets in particular.
1. Attribution: CC3 endorses whatever attribution a data providor can dream up. For most of the types of information CC3 is designed for, this is not an issue, but when data derived from 20 different sources is remixed, mashed up & put on a user defined online map, then it can become very difficult & sometimes impossible to ensure the attribution requirements are met.
As some examples of how this might happen, see: http://www.urbanresearchmaps.org/oasis/map.aspx http://eie.cos.gmu.edu/WMSUniPortal/ http://www.exploreaustralia.net.au/Queensland/Whitsunday-Islands
All these sites offer a range of datasets available as map layers, or text/images on screen. These are generally sourced from a wide range of private, government & NGO organisations. Applying arbitrary requirements for attribution for each dataset can make it very difficult to establish sites like these, yet enabling these sorts of data mashups is probably the main purpose of freeing up data, and how the greatest value of the data can be realised.
Tabular data, for example, summaries derived from LINZ, StatsNZ, DoC, MFish, MFE, NIWA, Landcare & GNS data may not be possible if they depend on each sources' attribution requirements to be complied with, especially web based tools allowing such processes & analyses to be controlled by the user (eg: Web Processing Service: http://www.opengeospatial.org/standards/wps) and not the provider.
At the very least, NZGOAL should be suggesting a standard, faciliative attribution requirement that can be reused just as CC3 itself is intended to be, or, preferably, NZGOAL should recognise that attribution is about branding, and should play no part in the provision of open government data.
2. Provenance. CC3 has nothing about data provenance or metadata. Without information about a datset, describing it's reliability, accuracy, precision, dates it represents, etc, making any given dataset freely available is of very limited merit, as such metadata is critical for users to make sensible decisions about what a dataset is appropriate for. For example, releasing a road centreline dataset is very useful, but only if we know when it was current, what sort of accuracy & errors it has, etc. It two or three such datasets are released, without such information, which one should be used where? Such data can not only become inaccurate, but actually misleading.
While I believe the move to CC3 is positive for most types of information, it was not really designed for open datasets, and using it for this purpose has real shortcomings that I hope NZGOAL will look to address.
NZ is party to the ANZLIC metadata profile, & I suggest that NZGOAL identify a suitable, standard profile which Govt agencies should be expected to comply with when data is released for public use.
