Book Project: Web 2.0

From IRF Wiki

Web 2.0 and the Semantic Web

• The questions raised at the IR Forum are discussed in the context of the patterns.
• Further ideas and research are added from the research interviews.

Ideas from the proceedings

The Scope of the Semantic Web: The promise of the semantic web was met by enthusiasm by some at the conference but strong resistance by others who argued that the idea represented the return of over-ambitious ideas that were tried and failed by previous generations of artificial intelligence reserachers. In this sense, those arguing against the semantic web were identifying a scope error.

The Promise of Web Services: One participant at the conference rejected the notion of searching the web and finding a web service and then being able to use it. This would be impossible because most web services would have to be documented to be understood, and such documentation is beyond the scope of the machine readable descriptions of web service interfaces.

Perpetual Beta: One of the most interesting concepts to come out of Web 2.0 is the idea of launching software when it is usable but before it is complete. This notion of the perpetual beta is closely related to the principles of agile development because it acknowledges that the best way to improve the software is to gain experience using it. The value of the experience outweighs the frustration anyone may have about the unfinished nature of the software.

What are Web 2.0 and the Semantic Web? Definition(s)? Competing views. Should the goal be an AI-rich Web - computers made to be more human-like and able to interact with each other more or less autonomously - or one of computers doing what they do best and thereby acting as tools that can aid people in doing what they continue to do best?

• Agreement on how “Web is evolving from being an information dissemination system to something like an transaction platform to ... be more of an interaction and collaboration platform,” thanks largely to the “democratization of innovation, which is becoming possible with a host of technologies and models around Web 2.0, where we are able to actually harness collective intelligence from the users”
• enhanced data and plug-and-play of data to achieve new things.
• web content and the functionalities associated with it have to be available in multiplicity of devices and forms.
• Semantic Web is “ simply a way of structuring information in a way that it's processable by more than one application. ... need to add a level of semantics into that, so that one can represent the constructs that were being used by the application and then one can start to share those. ”

A fundamental design and economics issue between:

• a lead design - traditional software engineering and structures and
• democratized design - community-based mashups, etc.

Key idea and trend: Consumers driving innovation (Or are they?):

• Communities on the Web - Korea
• Mashups
• Masses voting on new technologies

New, agile software development techniques and structures - new conceptual underpinnings and new technical architecture:

• light-weight programming modules -- RSS, RDF, or Ajax,
• Software as a service and the issue of multi-tenancy (??)
• Mash-ups - building new apps from old ones, hybrid apps, fusion of apps and data sources. Mash architectures from Intel.
• accelerating away from the notion of software and systems as engineered; new model is based on improvisation, iteration and interaction-driven design. “What was good engineering is now perceived as over-engineering” and waste.
• “ ... the scriptification of software and the scriptification of process design, i.e. Pearl, Python”

Implications of new, more agile SW development:

• a new notion of what a process is; SAP is a hard, well defined, very rigorous process, but now seeing more and more soft processes. Definition: 'easily modifiable', inexpensive to modify, bootleg services and processes in organizations.
• Largely driven by employees’ personal experience out on Web: Increasingly, people in the business enterprise are looking outside and say, hey, if I can google things externally, why can't I google them with a structured and semi-structured information inside the organization? How do enterprises import offerings in ways that may bypass or disintermediate the ERP - very important issue.

Web 2.0, mash-ups, etc. require improved data/knowledge representation:

• a lot has to happen on the technical front in terms of expression of meaning, knowledge implementation, business rules that support automated reasoning, intelligent agents that actually allow for this autonomous exchange of information, digital signatures that will allow me to authenticate and verify the identity of groups that I'm seeking data from.
• Information format standards - for structured and unstructured documents - are increasingly important - consistency of data becomes very, very important.
• Improved meta data appears to be critical, but will this be drive by traditional industry consortia efforts or by customers themselves? Can it be automated, or does it need to be done “by hand” as software gets coded and databases designed, for instance?
• Likewise, taxonomies - or folksonomies - and ontologies.
  • AI may be useful to handle non-tagged data
  • Semantic Web has best chance of catching on in new sites, where semantics are built in by design.
• Some community semantics efforts useless - e.g. Flickr
• Tag-searches limited because of simplistic string matching algorithms - one typo and no match.
• Industry-led efforts promising, especially in paid content industry: Bertelsmann/Sony indexing own music to help customers search
• Analytics-based search will become more important, which will allow people to get relevant information and high quality information.
• one of the major components of the semantic web is ontology ... a lot of work that needs to be done and it can not be done by one organization, one business or industry, it has to be done in a distributed manner and there has to be technology to integrate.
• Security is key, too: How to verify the identity and accuracy of data sources? Web services aren’t as secure as they need to be, yet semantic web applications are being built on top of those services.
• Management of intellectual property rights.

Stream-based data:

• Increasingly important, from sensor networks, for example
• New paradigm: Aim is not to store data but results of computations on the data
• Not easily addressable with traditional RDBMS

Three markets for new Web platform, each with own requirements:

• consumers
• small businesses
• large enterprises

Book Chapter: [1]