Ian Forrester's lifestream

In this session I’ll demonstrate and explain a method to use socially authored data as a proxy for currency and as an authoritative information source. This data can be used to provide context and add further contextually relevant content and meta-data to newly published and pre-existing media archives providing new ways to navigate the content.

By looking at how we can use sites like Wikipedia, Freebase and DBpedia as authoritative sources of content and meta-data and utilising services such as del.icio.us to provide a measure of popularity and currency (what is being discussed at this point in time), we can generate additional meta-data that can be used to provide new routes through existing content archives that start to help support a wilfing based browsing approach.

These techniques will be demonstrated via a working prototype that was developed as a research project for the BBC, targeting the main BBC News site, with the aim of augmenting the existing archive and providing different routes through the content.

I’ll examine some of the problems inherent in the method, such as contextual ambiguity (Australia the country vs. Australia the song), bias in the data-sources (e.g. the sheer volume of different tags for Google in del.icio.us) and how some content can be linguistically, culturally or geographically distinct and thus less relevant.

In this session I’ll demonstrate and explain a method to use socially authored data as a proxy for currency and as an authoritative information source. This data can be used to provide context and add further contextually relevant content and meta-data to newly published and pre-existing media archives providing new ways to navigate the content.

By looking at how we can use sites like Wikipedia, Freebase and DBpedia as authoritative sources of content and meta-data and utilising services such as del.icio.us to provide a measure of popularity and currency (what is being discussed at this point in time), we can generate additional meta-data that can be used to provide new routes through existing content archives that start to help support a wilfing based browsing approach.

These techniques will be demonstrated via a working prototype that was developed as a research project for the BBC, targeting the main BBC News site, with the aim of augmenting the existing archive and providing different routes through the content.

I’ll examine some of the problems inherent in the method, such as contextual ambiguity (Australia the country vs. Australia the song), bias in the data-sources (e.g. the sheer volume of different tags for Google in del.icio.us) and how some content can be linguistically, culturally or geographically distinct and thus less relevant.

• Clean separation of functionnal (what the application is doing) and non functionnal (e.g look and feel) aspects.
• Independence from devices and execution platforms: no vendor, device or platform locking.
• User friendlyness : no need to be a coder.

Open Mashups provides answers to all these requirements by applying in an open web context rigorous formal techniques from the MDA /MDE (Model Driven Architecture/Engineering) world. We defined a dedicated meta-model that represents an application from a very functionnal and declarative standpoint. This meta-model is the cornerstone of our approach: we provide an easy to use graphical editing environnement to depict application models (both for the GUI and the behavior of the application), and write code generators that produce executable code from these models. This approach allows to hide the complexity to the mashup creator that only needs to understand and manipulate a limited number of concepts, relying on code generator implementors to take care of platform and device adaptation.

The Open Mashups editor is implemented as an Firefox extension that also include a runtime engine to test applications on the fly. This component will soon be provided as an open source software. Code generators are implemented using QVT model transformations.

We’ll demonstrate the end to end toolchain, from application creation to code generation, leading to the same application running on a mobile phone, on a web page or as a desktop widget.

• Clean separation of functionnal (what the application is doing) and non functionnal (e.g look and feel) aspects.
• Independence from devices and execution platforms: no vendor, device or platform locking.
• User friendlyness : no need to be a coder.

Open Mashups provides answers to all these requirements by applying in an open web context rigorous formal techniques from the MDA /MDE (Model Driven Architecture/Engineering) world. We defined a dedicated meta-model that represents an application from a very functionnal and declarative standpoint. This meta-model is the cornerstone of our approach: we provide an easy to use graphical editing environnement to depict application models (both for the GUI and the behavior of the application), and write code generators that produce executable code from these models. This approach allows to hide the complexity to the mashup creator that only needs to understand and manipulate a limited number of concepts, relying on code generator implementors to take care of platform and device adaptation.

The Open Mashups editor is implemented as an Firefox extension that also include a runtime engine to test applications on the fly. This component will soon be provided as an open source software. Code generators are implemented using QVT model transformations.

We’ll demonstrate the end to end toolchain, from application creation to code generation, leading to the same application running on a mobile phone, on a web page or as a desktop widget.

The current ubiquitous Web is a “Web of Documents” where documents (Web pages) are connected by embedded hyperlinks (URLs or Uniform Resource Locators). The popularity of the “Web of Documents” (or “Document Web”) sometimes obscures the fact that from the onset Tim Berners-Lee envisaged a broader “Web of Data”. In this broader “Data Web”, data items are uniquely identified by Uniform Resource Identifiers (URIs), with these data items connected by typed links conveying implicit meaning (semantics) resulting in sparse or dense clusters of “Linked Data.”

Only recently have people started to understand the broader “Web of Data”, due to users demanding their right to reuse their information elsewhere. Termed “data portability”, the ability to store information in a format which could be rendered simply and easily is the one key into the “Web of Data”. However, “Linked Data” is essential for such a web to exist.

There are several questions that must be answered:

• What exactly is Linked Data and how does it fit into the Web of Data?
• How do we develop and use Linked Data?
• How do we deploy it, and get maximum efficiency from those links?
• How do we avoid problems with data access and unambiguous naming?
• How do we avoid problems with data reference and ambiguous association?
• How does Linked Data relate to Data Space philosophy?

This presentation will investigate the above issues and discuss a range of Semantic Web and data portability technologies. There will also be an explanation of how Linked Data works inside OpenLink Software, Virtuoso, and ODS.

The current ubiquitous Web is a “Web of Documents” where documents (Web pages) are connected by embedded hyperlinks (URLs or Uniform Resource Locators). The popularity of the “Web of Documents” (or “Document Web”) sometimes obscures the fact that from the onset Tim Berners-Lee envisaged a broader “Web of Data”. In this broader “Data Web”, data items are uniquely identified by Uniform Resource Identifiers (URIs), with these data items connected by typed links conveying implicit meaning (semantics) resulting in sparse or dense clusters of “Linked Data.”

Only recently have people started to understand the broader “Web of Data”, due to users demanding their right to reuse their information elsewhere. Termed “data portability”, the ability to store information in a format which could be rendered simply and easily is the one key into the “Web of Data”. However, “Linked Data” is essential for such a web to exist.

There are several questions that must be answered:

• What exactly is Linked Data and how does it fit into the Web of Data?
• How do we develop and use Linked Data?
• How do we deploy it, and get maximum efficiency from those links?
• How do we avoid problems with data access and unambiguous naming?
• How do we avoid problems with data reference and ambiguous association?
• How does Linked Data relate to Data Space philosophy?

This presentation will investigate the above issues and discuss a range of Semantic Web and data portability technologies. There will also be an explanation of how Linked Data works inside OpenLink Software, Virtuoso, and ODS.