Meshing Information and Mashing Functionality
Linked Data and Mashups rely on open standards to integrate information and furnish collaborative features, but differ on their uses and the intricacies of combining disparate data and features to build new functionality from cloud resources. Both provide emerging standards to provide content and shared applications for readers/users from existing Web resources coupled with proprietary enterprise data. Each promises increased productivity by combining and efficiently repurposing content and features for focused tasks and specific audiences.
Linked Data employs basic identification of content to loosely mesh and deliver shared semantic content of raw, personalized data from disparate sources. Likewise, Mashups employ semantic tagging but require logic through coding or high-level scripting using open APIs to provide functionality from disparate cloud services melded with proprietary data to render a new knowledge set or services. While both hold great promise for Web 2.0 moving to Web 3.0 strategies, each has significant differences in implementation and usage. Both technologies remain in their infancy to date but they are growing up fast.
This posting attempts to confront the similarities and differences of these technologies and their emerging standards while providing use cases and current examples. My intent in this and following postings is to highlight the promise of each technology to help developers and information developers catch a glimpse of a possible new frontier before making catastrophic decisions using yesterday’s technology. I don’t want anyone to pour concrete now on the plot they plan to garden next spring.
Quick Note: This posting comes from comments presented by Kingsley Idehen, President and CEO of OpenLink Software and my religious reading of Dion Hinchcliffe and Jeff Hanson. All commented on my About Mashups and Linked Data posting.
Linked Data and Mashups: Comparison and Contrasts
Linked Data and Mashups see the Web as a collection of objects of raw data and open functions rather than formatted documents residing on websites or proprietary applications. Both employ open Web standards to consolidate content and coordinate tasks using open protocols. Both promise quick and easy development for enterprise and commercial products and publications without in-depth development costs or expensive, proprietary tools. And both utilize structured markup as metadata to identify information based on its semantic meaning. Essentially, Linked Data and Mashups expose, share, and connect data to render comprehensive knowledge and actuate collaborative, complementary program features from the Web cloud.
Linked Data employs simple HTTP protocols to connect exposed data stores and stream unedited content quickly to interested readers before publishing as contextual knowledge. As a subset of the Semantic Web, Linked Data relies on semantic markup to define the meaning of content and then employing dereferenceable URIs (semantics) to locate and deliver Web content as URL addresses. It is Tim Berners-Lee’s vision of the Web as a universal data, information, and knowledge exchange.
The premise of Linked Data is this: The more connected data you have, the more powerful the information and more accessible the knowledge. It is the meshing of information without relying on published documents or formal web sites. Its promise is to deliver personalized content to each individual based on his or her unique needs. As Kingsley Idehen playfully puts it, “Linked Data basically delivers the ability to Mesh disparate data sources rather that settling for brute-force data mashing as exemplified by Mashups.”
In contrast, Mashups are more complex while promising more functionality, risk, and coding effort. Mashups require advanced logic to build consolidated applications using AJAX (Asynchronous JavaScript and XML), REST (representation state transfer), RSS/Atom feeds, gadgets and widgets that incorporate new processes and integrate with Enterprise 2.0 services.
Both emerging methodologies provide a glimpse of the future of programming and information development. As Dion Hinchcliffe writes, “Web 2.0 [provides] alternatives at a fraction of the cost of their enterprise-class predecessors, even if they don’t have exactly the same functionality. Enterprise 2.0 solutions, depending on their feature set, are becoming candidates to replace existing document management and knowledge management systems, enterprise portals, and even larger enterprise suites such as CRM systems.”
Problems with Employing Linked Data and Mashups
There are problems here that need to be confronted by proponents of both technologies:
- Data not ready to be shared. Much of the data in the world is not structured to be accessed and shared. Governments, enterprises, and social groups are not interacting together to provide access to their data. You can scrape data from FaceBook, but not MySpace, or Twitter, or Flickr, or YouTube. Each is its own silo. Tim Berners-Lee calls this “Database Hugging.” And let’s not explore the massive data tied up in corporations and governments.
Unique URIs are needed to expose linked data and a hierarchical ontology for each domain. Some of this work is being done at dbpedia.org and linkedgeodata.org and elsewhere, but much more work needs to be done to create and import ontologies and define standards. - Some database hugging is reasonable. Legitimate reasons for holding on to proprietary data exist. Medical and financial information needs to be private. Internet service providers better not be handing out my personal information. And companies have good reason for keeping information private, especially when it would be equally valuable to their competitors and less valuable to the public. Sometimes keeping information far from the Web is the responsible and legally-wise policy.
- Throw away database queries and applications of the last 20 years? Are we really supposed to throw away decades of database research designed to store, index, and query datasets to give way to open and accessible text files over the web? Combining SQL queries for legacy databases and common semantic tagging needs to be adopted.
I acknowledge that there is no reason to rush headlong into the open display of information and application functionality at this time, but there are scenarios where it makes sense now.
Immediate Uses of Linked Data and Mashups
While much has to be worked out in adopting Linked Data and Mashups, now is the time to determine if these technologies and methodologies are ripe to exploit in some fundamental way right now:
- Scientific research needs Linked Data now. I was offered to apply to position as a writer for a leading-edge nanotechnology and DNA research company. They wanted someone to handle content to share between scientists both internally and externally. My input as an experienced information developer was to establish an ontology of common semantics to automate content markup based on semantic to facilitate the interaction between research teams and provide real-time data. They didn’t need to throw this data over the wall to a writer to publish on a web site or through formal papers. They are getting back to me.
Example: http://bio2rdf.org/–Semantic web atlas of post-genomic knowledge. - Shotgun communication. Companies wrestle with all of their diverse content when releasing products and service from the different R&D, technical support, marketing, and writing teams in individual silos. All of this overlapping information needs to be brought together to identify cost savings and get the right information in front of the various types of customers (prospective customers, customers needing best practices, upsell customers, internal employees, et al).
Example: See my posting on this issue and an example of shotgun communication. - Low-hanging opportunities. Some mashup applications are easier to implement than others. These high-value, low-cost implementations need to be taken to the market to build on for more complex and productive products.
Example: Housingmaps.com.
Also refer to Enterprise Web 2.0 postings from Dion Hinchcliffe for a more authoritative validation and some additional best practices.
My Direction in using Linked Data and Mashups
In my next posting, I will define a project and take it to its logical conclusion while sharing my travails and successes. I will provide a product requirements document to identify my intent and define success. I will then share my findings. I plan to test out my previous postings about linked data and mashups, the myth of single-source authoring and publication, shotgun communication, and employing the confluence of content on the web, as well as the experience and expertise of real subject experts. I am going to hold my own feet to the fire.
November 24, 2009
Tags: Linked data, Mashups, Semantic Web, single-source authoring Posted in: Cloud Computing, Information management, Knowledge management, Linked data, Mashups, Ontologies, Semantic Web
2 Responses
uberVU - social comments - November 25, 2009
Social comments and analytics for this post…
This post was mentioned on Twitter by shaungamboa: #semantic_web Meshing Information and Mashing Functionality | Mashstream: Linked Data and Mashups .. http://bit.ly/4xRW0S…
Dear Author mashstream.com !
Just that is necessary. An interesting theme, I will participate.
Leave a Reply