[Aust-NZ] Inquiry into Improving Access to Victorian Public Sector Information and Data

[Rob Atkinson]

These are key points I think the FOSS community should debate and
reach a consensus (if possible) on...


On Thu, Aug 7, 2008 at 1:34 PM, Robert Coup
<robert.coup at onetrackmind.co.nz> wrote:
> Hi Rob,
>
> Good points - comments/questions inline...
>
> On Thu, Aug 7, 2008 at 2:35 PM, Rob Atkinson <robatkinson101 at gmail.com>
> wrote:
>>
>> 1) There should be a general position regarding accountability, like
>> the intention (but not current practice!) of FOI laws, that taxpayer
>> funded activities should support free and open access for
>> non-commercial use to relevant information. This includes all spatial
>> information, with the well understood exceptions of personal private
>> details and national security. Specifically there should be a
>> presumption of public interest, with only commercial exploitation
>> restricted.
>
> What do you mean by commercial exploitation here? If BigCo wants to resell
> free and publicly accessible data for $100K per user per year and can find
> suckers to sign up for that, whats the problem?
>


IMHO there is a danger (maybe a certainty) that we wont get legitimate
access to data that is currently also sold commercially.  If we
disallow resale for commercial purposes, then public-interest access
can be supported without changing all existing commercial
arrangements.  I dont we think we really care how the cost is
apportioned within governments either, just that non-commercial uses
(public interest, researc, non-commercial software development and
testing, education etc) should not be prejudiced by any potential
commercial value of the data.

>> 2) There should not be a technical cost associated in accessing data -
>> i.e. it should not be bound to a proprietary on in-house custom
>> technology.  FOSS has an obvious role in providing a baseline for what
>> is thus acceptable - there should be an onus on data access methods to
>> provide a FOSS reference implementation, and this should apply
>> automatically. Critically we mustnt contemplate building private data
>> distribution arrangements without a commensurate capability to make
>> the same data visible and accessible using open standards and
>> licenses.
>
> Standards are the key things here imo, if the data is available in standard
> formats or through standard service interfaces then it doesn't matter what
> software is in use to deliver it. Whether governments should be
> using/preferring Open Source Software is a different topic from providing
> open access to public information. I'd much prefer data released today in
> any format, than in five years via a standard format. Governments (and all
> of us) should be striving to push standards but getting data out there is
> more important than getting it in the "right" form.
>

Agreed - for example a FOSS reference implementation that reads an
ESRI shapefile would be acceptable - to the extent that the shapefile
meta model  (aka Simple Features Level 0 Profile) is useful for data
transfer of anything but cartographic data products.

Open standards will be required for many data products, because there
is no de-facto proprietary standard that meets metadata needs. Even if
you ship boundaries as shapefiles, a protected area for example has
metadata that wont fit into a single database column. This needs to be
accessible too, as part of any real data product. We should plan for
this capability.

I also think the stronger requirement that there is a test against
FOSS reference implementation is valuable in practice. I've seen too
many "open standard conformance claiming" proprietary implementations
that just arent usable. With much more rigorous open standards
conformance capabilities in place this could be relaxed,

Rob (or was that B1?)

> Looks like Victoria is heading in healthy directions though :)
>
> Other Rob.
>
>