Public Sector Information handling and rights (Vic) : 1
Posted by Dave Bath on 2009-06-29
Let’s start going into some of my detailed thoughts about the various recommendations for improving access to Victorian public sector information (PSI) and data produced by a Victorian Parliamentary Committee.
While I’m writing specifically in a Victorian context, any arguments and observations can be applied to any jurisdiction, at any level, in Australia.
1. That the Victorian Government release a public statement indicating that it endorses open access as the default position for the management of its public sector information.
This is pretty simple, and it’s hard to think of anyone arguing otherwise. The tricky bit comes with what exactly is meant by "open access", and how loosely defined are the rules that allow exceptions from the default position.
But… are they including information collected in the first instance by councils? By outsourced service providers carrying out work on behalf of government? If not, why not? (More on this in future posts on recommendations that are further down the list)
2. That the Victorian Government develop a whole-of-government Information Management Framework (IMF) with the following key features:
- that the object of the IMF is to promote and facilitate increased access to and re-use of Victorian public sector information (PSI) by government, citizens, and businesses;
- that the default position of the IMF be that all PSI is made available;
- that the IMF define and describe criteria under which access to PSI may be restricted, or released under licence;
- that PSI made available under the IMF be priced at no cost or marginal cost; and
- that the IMF establish a systematic and consistent whole-of- government methodology for categorisation, storage and management of PSI.
Let’s not worry about duplication between recommendations 1 and 2 from the committee. Let’s not worry about the "Information Management Framework" being desirable, and it’s contents as described.
What we should worry about is that most of the dot points in recommendation 2 shouldn’t need to be stated: these were supposed to be in place decades ago. Indeed, you could argue that various statutes with "Archives", "RecordKeeping" and the like in their titles already demand that such a framework be in place, and that it is implemented consistently across the whole of government.
For example, we do have a whole-of-government methodology for categorization, storage and management of PSI. Recordkeeping requirements federally and at state level are no different, so it’s worth looking at what National Archives Australia (naa.gov.au) has to say about what should be in an information management framework.
Among the things you must know about an information item are:
- What is the subject matter? This relies on a classification system, and could be a very-fine-grained set of Dewey numbers, or could be a government thesaurus.
- Who is the author/custodian? This could be an individual, a role, an agency, etc
- What is the audience/visibility? This includes sensitivity (secret, commercial-in-confidence, etc – with "undefined" really implying that it should be freely available to all) and who reads the document (the general public, for example).
- Might the information require special handling (e.g. if it relates to evidence in criminal investigations).
The problem is, despite these requirements being around for decades (well before computers were invented), there is, in Victoria at least, no general understanding of the requirements or available toolsets, or worse, antipathy towards the satisfaction of those requirements. Why else would the Victorian Auditor General have slammed recordkeeping awareness and practices throughout the Victorian government? (See "Vic Gov Recordkeeping Slammed By Auditor" (2008-03-21).
As the auditor said:
There is a lack of understanding by senior management of what good records management looks like and, therefore, what should be done.
I’d say that too often, good records management is seen as an unnecessary expense by many operational areas, or good records management is seen as limiting the options of those decision makers who do not welcome close scrutiny.
The "unnecessary expense" excuse is a bit like someone saying "I don’t need to record petrol I put into my car, because that record does not help me keep my car going on a day-to-day basis". True, but without those records, you risk poor management (more petrol is required as cars become inadequately maintained, you might be able to optimize routes, it’s hard to pin down rorting).
The key is classification. It’s no use having millions of books, for example, if you cannot get to the ones you want, which is why when a book comes into a library, it is given a Dewey number, and entries made into indices by author, subject/keywords, and title. Years ago, we did this will bits of cardboard, placed into various catalogs. These days, with electronic databases, you only need to create a single record, as this single record can come up when querying from different perspectives.
The Australian Government Recordkeeping Metadata Standard (2008) describes these information items. The full list is as follows:
- Date Range
- Related Entity
- Change History
- Security Classification
- Security Caveat
- Integrity Check
- Document Form
These categories and the allowed values in them are defined by various encoding schemes and lists, including (at federal level) the Australian Government Locator Service AGLS) and (at Victorian Level) the government thesaurus (see Jan 2008 v2.4 editions as HTML or PDF).
I happen to have worked with one of the smart cookies who designed at least some of these documents, and I have great respect for that person, and the work they produced. Classifying a document using these schemes is pretty easy for anybody who can understand the contents of a document, and should be a no-brainer for the author(s) of a document.
So, why do so many government documents have such rubbish in them? How can you verify my claim that the metadata of these documents is almost unmaintained?
Simply go and grab any Adobe Acrobat file (PDF extension) from government, use the File/Properties menu, and see if the title, author, subject and keywords make sense when compared to the document. If these are not filled in, how can a catalog (of any description) help you find anything?
Let’s imagine something that we should be able to do. Imagine you are running Centrelink in an area with a large proportion of residents not having English as their primary language. You need to be able to get all of the documents for a particular language, with an audience of "citizen", from one or more agencies (including Centrelink, but also health agencies, agencies concerned with discrimination), relating to just a few topics, effective now or becoming effective in the next few months.
Getting the list of all of matching documents (and probably downloading them as a bundle, or ordering copies of all the pamphlets) should be trivial if your catalogs are anywhere near the required standard.
Again, I’ll stress that these sorts of things have been required since democratic government bureaucracies were invented. As recordkeeping and document management are so poor (in Victoria at least), why and who have we allowed the system to deteriorate relative to the information loads and the capabilities of technology?
It seems to me that find the answers to "why" and "who" would be essential to encourage better compliance and adequate performance.
It is also worth asking if state governments should be the custodians of the classification systems. Personally, I cannot see why each state should go to all the expense of developing such systems, when the same classification scheme can be equally applied across all Australian jurisdictions. What could possibly be the differences between states (apart from the enumerated list of values, like names of geographical regions and department names) that would prevent such a consolidation? More importantly, what intergovernmental inefficiencies, either for operation of existing agencies, or with restructuring of responsibilities between state and federal levels, would necessarily arise from disparate classification systems?
I’ll also remind readers (especially those in power), that the same level of records management is required for outsourced activities, where those activities relate to the work of government. Too often, I suspect that government functions are outsourced at least with part of the motivation being that those records will be difficult to track down, and the agencies will hope to wash their hands of ensuring the proper handling of that information.
3. That the Victorian Government prospectively apply the Information Management Framework to its public sector information.
This seems to be a cop-out, admitting that there has been no real information management discipline applied across all agencies. Being prospective, classifying documents as they are created and/or updated make some sense for efficiency, because it would be burdensome and not very effective to go through the blegabytes of information that is no longer applicable. However, surely there should be something saying "and within a reasonable time frame, all active documents produced in the last n years will also be classified appropriately.
I have an ugly suspicion that when "prospective" is used, it will be seen as only applying to new documents, rather than to documents that are updated. I hope I’m being far too cynical on this matter… but my experience tells me the cynicism is warranted.