Linking research & learning technologies through standards

SIF Association AU recently held a two day workshop for the Data Standards Working Group, which has been working for the past couple of years on the Australian data model and specification for SIF. These are some of the highlights of the meeting:

New SIF Association US Standard

SIF Implementation Specification 2.4 is going to be released in early June; a preview of the features to be included is already available. (See also Larry Fruth’s presentation (PPT) at the recent IDEA10 event.) The new release of SIF features new objects and attributes, including improved coverage of assessment and its alignment to curricula, and objects to support special programmes for staff and students (student participation, professional development). But there are two major additions in this version taking SIF in new directions.

• Instructional Services (see SIF Implementation Specification 2.3) are the means through which SIF supports teaching and learning. The use cases and objects supporting them have concentrated to date on assessment, and on modelling the curriculum that assessment is tied to (in American parlance, Learning Standards).

• SIF Association AU will be looking at applying Instructional Services to Australian use cases. The business context around assessment and learning content distribution is rather different in Australia and the US: some business analysis is needed to work out how to use the US services effectively, and how the objects should be tailored to Australia. Work is also needed to establish how the SIF Instructional Services fit in with standards with overlapping scope already established in Australia. Assessment use cases are the most straightforward place for this work to start.
• Readers of this blog will already know about the Semantic Web-based approached to curricula taken in the States by JES & Co. (Achievement Standards Network), and the interest expressed in Australia on exploring such approaches for the National Curriculum. JES & Co is already collaborating with SIF Association US, exporting their machine-readable curricula to SIF formats; so the SIF and ASN approaches to curriculum will be complimentary in Australia.
• New to SIF Implementation Specification 2.4 is the Content Catalog object: this will be used to model discovery of learning objects across multiple repositories and sources. The intention is not to supplant existing discovery protocols (such as SRU and OpenSearch) or metadata formats (such as LOM). Content Catalog queries are meant to act as a common, profiled search query, capturing generic instructional needs (including for instance alignment to Bloom’s Taxonomy); such queries can be farmed out to various sources connected via SIF, providing in effect a federated search protocol. The planned Content Catalog object is described further in a SIF presentation.
  • The proposed approach of federated search across a Zone Integration Server (ZIS, the SIF equivalent of an Enterprise Service Bus) is solving a clear problem in the US, where school districts are silo’d from each other and can benefit from preexisting SIF Infrastructure. There are other possible ways of implementing federated search in Australian education; some infrastructure to that end already exists in Australia (LORN, TLF), with broader scope than is typical of the US. So further analysis is needed to see how the SIF Content Catalog will integrate with current infrastructure: in the meantime SIF Association AU is prioritising assessment instead in its investigation of Instructional services.

• Also in SIF Implementation Specification 2.4, SIF is adding Zone Services. These provide a way of emulating Web Service behaviour within the existing SIF Infrastructure through exposed methods (including service input objects, service output objects, and notifications of when an application event has occurred). The addition of service-like behaviour to SIF will help expand messaging functionality beyond CRUD, and will forestall proliferating new data objects by using service message bodies instead for novel purposes.

  The intent is for SIF down the road to support full SOAP, alongside its existing messaging infrastructure: web services will form the basis of a second reference SIF Transport, specified with normative dependencies (SOAP, WSDL, WS-Addressing, MTOM, etc.) The two transports will be complementary; for example, the native SIF transport can be used within the school’s trust environment, and Web services can be used to communicate with external bodies. Web service applications of SIF will be forwards and backwards compatible with the native SIF applications, interacting through SIF agents, because the ZIS will support both transports, and will specify web service functionality through WSDL.

  Zone Services are a way of introducing service functionality early on to SIF, without forcing changes to existing infrastructure. The initial zone service instances will involve moving student transcripts between jurisdictions (cf. the Student Data Transfer Note in Australia); reporting of assessment between organisations; and tracking identifiers of students and staff between multiple jurisdictions (staff, student locator services).

Coordination between SIF profiles

SIF Association AU and SIF Association UK have pioneered applying the SIF Data Model to a new national context, and profiling the data model to meet local concerns. There is a concern that extensive localisation has become a problem for vendors, who now have to support three unaligned localisations of the same specification, each with its own versions. Updates in one localisation also need to be considered for inclusion in the others—typically US updates reflected in AU and UK—but there is starting to be feedback in the other direction as well. Without alignment, this process is becoming unwieldy. With the adoption of SIF in an increasing number of countries, this difficulty will only increase.

As a result, the SIF localisations are moving to align their conceptual models explicitly, identifying where they are using the same name for different concepts, or different names for the same concept. There is also a desire to decouple the data model of SIF from the infrastructure model, so that updates to infrastructure (e.g. the new Zone Services) can be taken up by localisations as globally applicable updates—independently of changes in data objects, which are often profile-specific.

The outcome of this will be a Global Trunk data model for SIF, which will be the common core of the various localisations; this means the SIF Association US data model will now also be counted as a localisation. A Special Interest Group is being assembled to investigate further how localisation can be made more effective, including obtaining tools to manage the common data structure and the mapping to different localisations; where possible, such mapping should be automated.

Codeset Standards

SIF relies critically on standard codesets to exchange data so that it can be commonly understood between systems. This is critical to vertical reporting from a range of jurisdictions, and much of the impetus for SIF Codesets comes from the requirements of vertical reporting. Hence the use in SIF of NCES codesets in the US, and ABS codesets in Australia.

The ABS is interested in promoting further consolidation and standardisation of codesets, and presented to the workshop on SDMX, an initiative building standards for the exchange of statistical information; SIF may consider make their codeset use in vertical reporting interoperable with SDMX.

There was also a presentation on METeOR, the Metadata Online Registry of the Australian Institute of Health and Welfare. The registry stores standard codesets used in health, housing and community services in Australia; it has a comprehensive search interface, and serves as a good example of what can be achieved when a sector applies strong coordination to its codeset use. That strong coordination, of course, is much easier to achieve when it is driven by a central mandate—as is required by vertical reporting.

Learnings from SIF Association AU work

The workshop provided an opportunity for the data standards working group, which has been working on realising an Australian SIF for the past couple of years, to reflect on what has been achieved to date, both in the process of formulating the SIF Implementation Specification (Australia), and in the pilots on SIF over the past year.

One of the key points raised is that the Australian pilots are doing pioneering work with several issues, such as authentication provisioning and multiple role-based authorisation. These do not seem to have been anywhere near as pressing concerns in the US and UK deployments of SIF, so that Australian pilots are having to confront them anew. The reason identified for this is the coverage of SIF Implementations in the different countries. In the US school districts are silo’d, and much of what they do is independent of the State as a coordinating agency—let alone other States or federal structures. In Australia on the other hand, SIF is a nationwide deployment, and data is routinely having to cross boundaries, not only between individual schools (e.g. teachers with different roles in different schools in the Tasmanian pilot), but also between different states (e.g. Triborders programme, tracking itinerant student movement between WA, SA and NT). The difficulties in successfully transferring data across boundaries have thus received more attention in the Australian deployment, and surmounting these is identified as a major benefit.

SIF in Australia, it should be added, has been applied to a large range of problems in the ongoing pilots, in conjunction with a variety of technologies. This has had several downstream benefits, as opportunities for further integration and interoperability have been identified. Identifying OpenID as a candidate technology for jurisdiction-wide authentication infrastructure, for example, has been invaluable.