posted on December 04, 2012 12:40
As those with even the briefest contacts with the Crossover Project will know, we're working on updating the research roadmap begun under the Crossroad project. Project partners and our animators are busy encouraging people to review the roadmap and comment on it. To achieve this, the various sections are presented on this site in 'commentable' form. We've done this using co-ment.com which offers a powerful commenting and text-editing platform. The software for co-ment is available free of charge and is open source, and the 'pro' version is not expensive.
In recent months there have been several discussions in the W3C eGov Interest Group concerning the Strategy Markup Language StratML. This is a serious attempt by various technologists in the US to create an open standard for this kind of thing. Since StratML is simply an XML dialect, and is being developed in an open space, it is an open standard and therefore meets the requirements of relevant US government policy.
I mentioned the Crossover Research Roadmap in an e-mail about our forthcoming Trans-Atlantic Research on Policy Modelling Workshop that I sent to several Washington-based members of the eGov Interest Group and Government Linked Data Working Group. Within 48 hours, Owen Ambur and Andre Cusson had converted the document into StratML and created an editable version on Hyperbase.com.
Impressive. Doubly impressive is that StratML is based entirely on W3C standards (XML and XForms).
But what is StratML? At its most basic, it's a pretty simple XML schema that can be used to publish strategy documents, including aims, objectives and stakeholders. See Gannon Dick's StratML version of NASA's strategy document for example. The hope is that strategy documents become linked and discoverable - you can see which agencies are targeting the same stakeholder groups, what the different approaches are to meet the same goals are and so on. Sounds like efficiency in the public sector - good. Let's look deeper.
Hmm… as well as some specific elements like
StrategicPlanCore it has elements like
Identifier. Looks a bit like Dublin Core to me.
StratML makes the very valid point that documents like this should be machine readable as well as human readable. The StratML way to achieve this is to offer an XSLT to create HTML documents from the XML - a perfectly valid way of doing things. Couldn't you do that with RDFa (or microdata) in HTML documents? Yes. Couldn't you do that by content negotiation between different representations of the same document? Yes. But that doesn't make it wrong to use XML & XSLT.
Where I do think StratML misses out though is on the linkage. If I look at the StratML version of the Crossover roadmap, I can see that, for example, Objective 1.8 has the identifier
_bdaf4222-3b23-11e2-b375-4576cf2fd515. This is a machine-generated identifier which is fine but it's specific to the environment in which it was generated. Greater use of linked data principles would enable the creation of individual identifiers that were more user-friendly and would exist independent of any document. Compare the StratML approach to, for example, the new European Legislation Identifiers, in which the use of URIs as identifiers and RDFa as the bridge between human and machine readability is more prevalent (see, for example, John Dann of the Luxembourg Government's presentation on ELIs in Paris 2012 (PDF).
Looking at StratML and ELIs, and being exposed regularly to what's happening 'over there' (the USA) compared with what's going on 'over here' (Europe) I think my biggest fear here is that Europe is embracing URIs as identifiers and the USA is focused much more on text wrapped in XML. The two mindsets don't always play well together and I worry that the Atlantic may be getting wider.
Even in an XML-centric world, URIs are surely the most appropriate identifiers for anyone wanting to work collaboratively online. These need to be designed carefully, backed by an intention to maintain them for the long term. From my point of view, agreeing on identifiers is a key to interoperability, no matter which technology is used. I've just concluded a study of current practice concerning persistent URIs. Many EU Member States have policies in place or are working on them and I'm hopeful that the report will be published soon.