About Mashups and Linked Data
In the past few postings, I presented some of the problems in managing information for corporate products and services using legacy authoring and publishing processes combined with new Web 2.0 communication practices. Take a look at these postings to get a clear view of where I’m coming from:
Shotgun Communication: Haphazard information distribution from disparate and insulated teams within an organization makes it hard for the customer to find usable knowledge.
Knowledge from the Cloud: The pros and cons of online information that questions information sources and their veracity. Not all information is created or maintained equally.
The Fog of Information: Too much information with too little functional knowledge permeates the Web.
The Information Confluence: Defines data vs. information vs. real knowledge.
These articles identify some of the basic problems of current information management strategies for all types and sizes of organizations. I will be adding to this list of failures and uncertainties of corporate knowledge management in the future. But for now, it’s time to turn the corner and start identifying some of the solutions to these problems. Let’s start with the basics.
In the next several postings, I will begin to identify new opportunities for capturing, presenting, and providing usable information with knowledge and features specific to each reader. This includes mashups, Semantic Web information, and linked data sources.
What is a Mashup?
First of all, we need to agree on a basic definition of a mashup. There are many types of mashups—commercial, enterprise, data, and knowledge mashups to name a few categories. But basically, all types of mashups consist of two or more unrelated sources to create a new entity as a product, service, e-book, or online application.
Mashups capture data and online applications from various points of origin and combine them to create a new functional entity specific to the needs of each corporation, reader, and customer. Mashups employ open APIs to create functional services that capture exposed data and application features from existing Web content and protocols to form a new type of online application.
One simple example of a mashup is an aggregated Sales application that integrates CRM and financial data with functionality from the Web and corporate backend data. This example mashup would employ real-time information, streaming content, and Web services to form a coordinated application using all of these data sources. Integrated sales information for the traveling sales person could be available from their smart phone or laptop. I imagine this sample application to include these integrated features:
- Streams real-time Web information of financial and customer relationship management (CRM) data from NetSuite or Salesforce.com, combining it with online maps to visually identify, locate and categorize customers for each geographical location. Using Google Maps or Mapquest APIs, each customer site appears on the map and allows the sales person to drill down using the map paradigm to identify customer sites to expose new sales or possible upsell opportunities.
- Background information and Request for Information (RFI) documents could be generated partly using semantically rich content from DBpedia, the semantically structured content from Wikipedia. Integrated and updated glossary definitions of domain vernacular, references to partners and competitors could come together as competitive analysis documents. Prospective customers could read marketing evaluations combined with general reference content, and links to trusted independent blogger opinions, all from a single document.
- Internal, proprietary customer data about installed products, contracts, and upsell possibilities can be integrated with the maps, reference information, and sales database to provide personalized content for customers.
Projections estimate that 80 percent of enterprise applications will be mashup applications in the future, with commercial applications and widgets already appearing on your desktop or favorite web sites. To accomplish this goal, open protocols and APIs such as Semantic Web markup, Service-Oriented Architecture (SOA), and the Real-time Streaming Protocol (RTSP), among other competing protocols currently being developed and adopted.
What is the Semantic Web?
The Semantic Web identifies content on the Web based on its meaning, providing intelligence from various data sources or actionable features from online Web applications. This is in contrast to “syntactic” markup, which only identifies the look and layout of content. Tim Berners-Lee’s, the inventor of the World Wide Web, is one of many champions of semantic web interaction as a consistent medium for the exchange of data, information, and knowledge (see the Information Confluence for the distinction). The Semantic Web relies on the markup of separate information fields to be understood by computers in order perform much of the perfunctory background operations involved in finding, sharing, and combining information from the Web to render as usable knowledge.
The Semantic Web continues to fall short of many of its highest ambitions to date, but is still supported by Tim Berners-Lee and World Wide Web Consortium (WWWC). Criticism about its practical feasibility, lagging progress, security and privacy, and need for additional markup has yet to be resolved. Regardless, many still see the semantic content as Web 3.0, the next evolution Web. It promises to provide intelligence and interaction for semantic publishing in scientific research, exposing experimental data as structured information for real-time sharing of information by researchers.
The Semantic Web is the only strategy proffered at this time to furnish intelligence and context across disparate Web information systems. Structured information based on context and meaning of information seems to be the only way to control Web data as it grows exponentially. Semantic publishing promises application interoperability and efficient, automated data integration.
One of the more realistic components of the Semantic Web is Linked Data, an environment that sees all entities on the Web as individual objects. These objects can then be intelligently combined and repurposed to create a new service, product, or knowledge set.
What is Linked Data?
Linked data simplifies much of the complexity of the Semantic Web. In contrast to the full Semantic Web, linked data publishes structured data using URI addresses rather than relying on a hierarchical (called ontological) cascading of parent/child relationships established by semantic markup. Using URIs, linked data handles everything on the Web as an object to be formed and presented as knowledge or actuated as new services and products.
The following diagram represents the standard concept behind Linked Data. Disparate data sources from data stores, documents, Web sites, and other cloud and internal data repositories come together organically to grow a new, fully-functional knowledge set, service, or product.
The main goals and concepts behind linked data includes these four principles:
- Use URIs to identify web objects.
- Use HTTP URIs to refer to and searched by readers and user agents from other applications.
- Provide useful information (e.g., a structured description or metadata) when the URI is referenced.
- Links to related URIs to expose data to improve searches of related information on the Web.
In my next posting I am going to take a stab at some practical uses of Linked Data applications.
November 16, 2009
Tags: Linked data, Mashups, Semantic Web Posted in: Linked data, Mashups, Semantic Web
10 Responses
uberVU - social comments - November 17, 2009
Social comments and analytics for this post…
This post was mentioned on Twitter by dhinchcliffe: Mashups and Linked Data: http://bit.ly/3hMrd2 Good primer with latest issues….
Nice post, but a few comments.
Linked Data is about using HTTP referencable Names (generic HTTP scheme URIs) for Data Objects / Data Items, that de-reference to HTTP accessible Data Sources (via Locator oriented HTTP scheme URIs commonly known as URLs).
There is a duality to the Generic HTTP scheme URI that enables Identity (Name)/ Access (Address) to exist in a single unit. Its this duality that enables the Linked Data magic whereby de-referencing a Data Objects Names results in access to the a structured data representation of its description (Metadata – a constellation of data relations that describe said Data Object).
Bearing in mind the above, Linked Data basically delivers the ability to Mesh disparate data sources rather that settling for brute-force data mashing as exemplified by Mashups
Enterprises have to look to work with disparate data sources in the the same way they work with data, hosted in a single DBMS from a single vendor i.e., they should look to JOIN structured Data en route to constructing holistic views over their data silos etc.. Of course, the same applies to the public Web since it too is proliferated with Data Silos courtesy of Web 2.0.
Kingsley Idehen
Kingsley: Thanks for the well-written clarification. I really appreciate you defining the differences between linked data (more meshing) and the brute force mashups. I will follow up on this with more research on a few of these issues.
Amazing as always
Twitted by Concepthubinc - November 18, 2009
[...] This post was Twitted by Concepthubinc [...]
Links for 2009-11-17 [del.icio.us] | Trinitude Network - November 19, 2009
[...] About Mashups and Linked Data | Mashstream Mashups and Linked Data: http://bit.ly/3hMrd2 Good primer with latest issues. [from http://twitter.com/dhinchcliffe/statuses/5796317657 [...]
Meshing Information and Mashing Functionality | Mashstream - November 24, 2009
[...] OpenLink Software and my religious reading of Dion Hinchcliffe and Jeff Hanson. All commented on my About Mashups and Linked Data [...]
Keep posting stuff like this i really like it.
Dear Author mashstream.com !
And I have faced it. We can communicate on this theme. Here or in PM.
The Knowledge Mashup | Mashstream - February 5, 2010
[...] the idea behind the Knowledge Mashup. Find the best pieces of information, tag that information, and then structure and sequence it for [...]
Leave a Reply