8. Evolution of Framework to an Optimal Architecture

What is the path to the future? Can such a framework stay current as technology changes?

Even more to the point, what if it turns out this is not the right architecture, or maybe even the right standard, to support semantic mediation and interoperability? We believe we have the right foundation for long-term progress, but many potentially game-changing technologies are introduced on a regular basis, so it is wise to consider those possibilities.

In this section, we go over the possible end-game scenarios for this architecture. In even the least positive outcomes, there are effective migration paths so that the work that has been performed is not lost, and can be easily used in future architectures.

We describe these scenarios starting with the least positive outcomes, even while believing the latter, more positive outcomes are most likely.

Worst-Case Scenario: Wrong Architecture

Even if this somehow turned out to be an ineffective architecture and paradigm, most artifacts and results can be easily re-used.

The current vocabularies that are used by these communities will have been well-described, in formats that can be understood by future generations even if they do not understand anything more than ASCII text and the English language.

The use of XML (and OWL or RDF) to capture these vocabularies and mappings means that their content can be readily transferred to other formats or architectures, using extremely common transformation approaches like XSLT.

The knowledge captured in the existing mappings can very likely be leveraged in producing next-generation data systems, in the same way English text files are being leveraged to produce structured data.

Even if the mapping approach is overcome by future approaches, the interfaces for transforming a local concept into a community concept may still be applicable, though the back-end service may operate in a different way.

Community Engagement and Best Practices

The involvement of the community in developing vocabularies and mappings is a necessary step to integrate the community's knowledge. As the community is engaged, they will simultaneously evaluate the value of the activity and its products, and assess whether there are better ways to achieve the same ends.

The practices we begin using now can be adjusted, enhanced, or replaced according to the evaluations of the community. Many of the technologies are largely independent of these practices, or can be adjusted to allow for other practices (e.g., remote telecons, or individuals working on maps rather than teams).

Meanwhile, a major accomplishment will be achieving community realization that each individual's way of referring to data results will not be enough for the community as a whole to be interoperable. This realization itself will lead to accelerated progress toward interoperable outcomes.

With this realization, the community will be able to make basic use of the work described here to perform fundamental alignment of concepts across different vocabularies. Even if not every application adopts the architectural framework, the ontology repositories will represent a store of community knowledge that will be important for future understanding of the systems.

Component Evolution and Best Practices

The components developed for this framework will be evaluated as work proceeds. It is our expectation that some components will be highly valued, others will need improvement, and some will be rejected in favor of different or newer solutions. As the architecture draws more users and applications, developers will provide additional and replacement capabilities for those initially produced. We support the development of such capabilities, in most cases eagerly and with our own contributions.

The architecture should not be dependent on the implementation approach of any single component, nor on a single provider of any given service, so long as its best practices (e.g., for referring to terms maintained by another entity) are agreed and followed.

Synergistic Development

We intend to present multiple demonstrations of interoperability with this Semantic Framework, developed by different members of the community. The simultaneous development of multiple applications and components by various participants maximizes enthusiasm, progress, interoperability, and robustness of the overall framework.

As the participants interact, identify issues, and come up with new scenarios, this will identify issues and validate many of the concepts in the Semantic Framework, and create a shared ownership that should drive it forward in optimal technical directions. The ability to demonstrate multiple applications leveraging the common framework will be a strong indication of the viability of the approach.

Best-Case Scenario: Semantic Web Integration

Ideally, the semantic framework described here will prove useful to the scientific community, and will beadopted to capture semantic meaning in a significant way. In this case, the applications will be able to provide additional interoperability to the science and related communities, and research capabilities will be significantly enhanced.

Beyond that outcome, however, the framework will also enable greater interoperability of systems and knowledge in the entire internet, as it grows toward an encompassing Semantic Web. This growth of semantic capabilities has been accelerating rapidly in the last few years, and will be increasingly a part of the interactions provided by the World Wide Web, in many cases invisibly to the users.

By choosing technologies which are likely to become a basis for the wider Semantic Web, this semantic framework enables the scientific communities to become engaged, visible, semantically useful contributors to the web itself. In turn this enables much broader appreciation of the science world by the general web community, with possibly far-reaching consequences.