Posted by cadsmith on August 10, 2007
When new platforms are envisioned, people tend to wonder what the killer app will be. In the case of the semantic web, where proof is structured in, a sweet spot for applications & services might be prediction. In general, some long-standing questions might have a mode where they remain asked over time, e.g. status type inquiries. These might have links to dependancies so that the answers need only be updated when the underlying significant data changes.
The status of the s-web itself at any particular time is also interesting. Built-in tests might describe how well the mechanism itself is working before it is actually used for other purposes. If one such web utility is down, there may be alternates available. Metadata for each instance of an SW would indicate its readiness & appropriateness at the time. There might be a management layer across these SW’s to handle the reporting & selections. They may not all be created alike, however, so results may change based on access to the content, & the state of the ontology, schema, rules, logic and so on. There may also be layers of users trying different approaches & experimenting with new techniques which may be outside of the most reliable choices used in the past.
If one had access to the results of the testing effort to date, then common questions such as “Is it done yet?”, “Does it work?”, “Is it fast?”, or “What’s wrong?” may be directed at the web. The system then reviews its general knowledge, results data, or user context & responds or asks for more specifics. Users trace why the system draws conclusions, & the same might occur the other way around to get a web-generated mix of qualitative & quantitative information. Answers then become customized to that effort, e.g.
- h/w & s/w were done, but OS has new version available, or
- 80% of the functions have been used okay, remainder unknown, or
- Seems to meet specs 90% of time, but need more memory to do better, or
- Passes diagnostics, but configuration has gotchas.
Additional annotations & opinions of testers would also become available as part of the answers. This system can then be polled prior to attending a product meeting where feedback would result in further staff tasking. Driving this from merely in-house use to tech support might add value if done correctly. Gaps in testing & analysis might be highlighted if these types of systems can interact at the methodological layers. An additional feature might be how to choose presentation, e.g. selecting type of graph & effective & most relevant data.
Beyond that, there is the possibility of development of further test-centric languages that recognize test-related ontology, tolerate multiple types of data nodes & agree on communication protocols, e.g. as digital radios do. It is convenient if the result data is stored in common format, interpretable by RDF-enabled services.
There may be new hardware components that are created to take the SW out of the research stages. This has precedent in the software industry, e.g. competition by the big economies involving the playstation, xbox, & wii. Also music semantics drive new players, & may lead to new types of instruments. SW device affinities might be along the lines of encryption, biometrics, storage, ISPs, translation, etc. Also the applications might concentrate on physical distribution, finance, disposable goods, food, medicines, infrastructure, etc. Various countries would emphasize industries that need shoring up.
Is there a future in semantic sports? Now is a good time for movements that cross boundaries. Perhaps one can ask the SW…
> Write the book on who looks good in version 3.
 “The Semantic Web as the apotheosis of annotation, but what are its semantics? “,
Yorick Wilks, 2005
Discusses the origins of the SW including AI knowledge representation, natural language processing, & trusted databases.
 “The semantic Web : a guide to the future of XML, Web services, and knowledge management” by Daconta, Oberst & Smith, 2003, ISBN 0-471-43257-1.
Talks about how businesses might apply SW capabilities.
 W3C Semantic Web Activity page:
 Wiki discussion: