Search Disabled |
|
|
|
Search Options | Help | Site Map | Cultivate Web Site | ||||
|
|
|||||
| Home | Current Issue | Index of Back Issues |
| Issue 5 Home | Editorial | Features | Regular Columns | News & Events | Misc. | ||
By Christian Guetl - October 2001
Christian Guetl discusses a proposal for improvements to get to a more structured and personalized World Wide Web. The aim of this article is to discuss a vision, an idea of a possible way of providing existing Web content with more structure and context, which may help users to get more relevant information from the Internet as well as automated software to collect and process task specific information.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
The strength of Integrated Framework for Augmentation and Enrichment of Knowledge (IFAEK) allows, unlike recent systems, the combination of additional information from more than one meta-service within the content of a particular Web document. Therefore individual meta-information from several sources related to users needs and their user profile can be provided. This will support and improve users identifying, browsing, managing and evaluating Web content, and therefore build a smarter Web. Furthermore, some advanced, improved and novel services, even commercial services, are made possible by IFAEK.
IFAEK may handle current and future client applications (e.g. PC, PDA, WAP phones, set-top boxes and hands-free browsing systems) and information services to build a more powerful net of information sources on the Web by the use of combination and integration of distributed information systems.
Since early days of the Internet, many have made comments about the problem of unstructured and passive information. To overcome this situation, consequently numberless ideas and research work has be done. For example, the Xanadu Project led by Ted Nelson, an early hypertext pioneer, has always worked hard on the ideas of a stable, structured and interactive system since the 1960s [25], also Herman Maurer and the Hyper-G team in late 1980s [23]. To give examples of some more recent ideas, Tim Bernes-Lee proposes the idea of an improved "Semantic Web" [4], XML workgroups of the W3C [28] are on the way to making the semantic Web a reality. However, at the time of writing this article, users still face a lot of problems. Some of them are itemised below:
C. Chislenko explained the current situation accurately: "The Web is probably the richest information repository in human history, but most of its information is passive and unstructured. The Web doesn't know what it carries and for what purpose, and the users cannot specify what they want from it." [6]
We believe that today techniques and pre-existing information services on the Web may already provide a first solution for a more structured and personalized Web. In the remainder of this article we will discuss the basic ideas behind this, argue why our vision could become reality, describe the framework IFAEK and show improved features and novel services.
Since late-1997, the author of this article has been considering the serious problems stated in the above Section, which resulted in a number of former research works ([9], [19] and [32]). Frequent observations of research works, standardization efforts as well as public and commercial services in a broad field of Internet information systems, related to this serious problem, show some isolated islands of improvements. This then prompts the author of this article to propose a novel Integrated Framework for Augmentation and Enrichment of Knowledge (IFAEK) to the existing document structure by collaborative and distributed work of individuals and communities.
The idea is as obvious as it is simple. On the one hand, there is an unstructured and huge information repository on the Internet. On the other hand, there are a lot of useful but isolated services there. Why not to use pre-existing information services and therefore exploit a great mass of human knowledge stored on the Internet? Why not to let pre-existing results of computer processes taken into account? Why not to use the knowledge of the huge amount of users and exploit their behavior? Why not combine all of them and therefore provide users with more relevant information? Why not to allow commercial services to be a part of it and enable new commercial services? The answer is a simple to use a smart framework (see below) combining information services and providing composite information in a quite personalized way for users. The valuable and novel point is that information from more than one meta-service dependent on users' needs can be managed and provided in combination with the corresponding content to a wide range of clients. In addition, active information contributions as well as the behaviors of users also have to be taken into account.
That simple solution, stated above, led us to emphasize some important requirements for such a smart framework:
(1) Users (other than the author) and groups of users must be able to create and manage their own sets of metadata and structures for their specific purposes.
(2) This implicitly requires that metadata must be manageable in a distributed environment (independent of the related document).
(3) Different sets of Quality metadata - dependent on the special needs - must be provided for the users.
(4) Existing geographical distributed metadata (services) must be easy detectable and selectable by the users. Both requirements are only partially fulfilled at the time of writing.
(5) A variety of pre-existing and new services have to be manageable to meet users needs. A proper framework has to provide mechanism to also integrate commercial services. That means that a kind of micro payment has been taken into account.
IFAEK may be one possible way to provide a generic solution. It enables the integration and collaboration of pre-existing and future information services (combining isolated solutions) by an open framework. Bearing in mind these ideas, the author of this paper has observed research work, standardization efforts as well as public and commercial services in a broad range of Internet information systems. These observations lead to some reasons why the IFAEK may be possible, which will be discussed in the following Section.
For example, existing standards and auspicious drafts from the XML family of technologies for structuring content on the Web [28] and SOAP, a protocol for exchange of information in a decentralized, distributed environment [29] are available. RDF, the Resource Description Framework provides Web-based metadata activities including sitemaps, content ratings, stream channel definitions, search engine data collections (Web crawling), digital library collections, and distributed authoring [30]. Another interesting item is the topic map subject [27]. On the other hand, interesting middle ware standards allow applications compiling distributed services by using ".net" [12], "Sun One" [26] and "mono::" [24].
For example XML tools [28], SOAP protocol implementation [10], JavaTM Remote Method Invocation (RMI) [20], ".net" [12], "Sun One" [26] as well as "mono::" [24] and topic map tools [27] may support the idea of IFAEK.
Within the last years, the information technology (IT) revolution has caused favorable trends in network and hardware techniques, which can perfectly support the proposed framework. Faster network links and connections as well as flat rate offers, internet-capable handheld phones and Personal Digital Assistant (PDA) facilitate nearly permanent access to Web content. The situation may positively influence the knowledge sharing process, the enrichment of information and meta-knowledge. Consequently, users are increasingly going to use meta-services to pinpoint relevant information as well as passively enrich information by their behaviors and by actively contributing their own knowledge.
A large and increasing number of users, belonging to manifold social groups, interests and being located in virtually every country of the world, produces and uses the content on the Web. [16] [33] The critical mass of users required to build a global collective knowledge base, which can be managed by self-organized meta-services seems to have been reached.
One example of a general metadata service is described in the research work of Gronbaek et al. The authors discussed an approach to utilize open hypermedia structures (e.g. links, annotations, etc.) as metadata. They proposed a general open hypermedia system, which can manage the metadata for arbitrary Web sites and allows collaborative work. The authors introduced an XML-based data format, the Open Hypermedia Interchange Format (OHIF), for the handling of such structures. [17] Further services like subject gateways, annotated links lists, etc. are discussed in [31].
Examples in the field of distributed metadata application are [22], [15] and [18]. An example of a pre-existing commercial service is Alexa [2].
On one hand, the huge amount of information available on the Internet cannot be processed and managed by human efforts. On the other hand, only automated computer processes - recent techniques are mainly statistic-based processes - do not meet the requirements for processing and managing mankind's knowledge repository on the Internet. Dramatical improvements can be gained by combining the intelligence and creativity of humans with the power of computers [1], [5], [6], [7].
It is well known in social science and business economics that people aspire to tribute within their social environment, their business and informal communities. In the Internet environment, it may be that people work on open source projects or run a homepage providing interesting information. Mark Frauenfelder discusses this effect in his article in the WIRED magazine [14]. The author stated the effect that people in Internet communities work just for "ego gratification". For example, they answer questions for other Internet users [3], or rate Internet pages for a community [11]. Frauenfelder notes the "egoboo" effect (short for ego boost effect), which is described by "the rush you get when you see your name in print [...]". Exactly these effects can be and already are exploited by services and start-ups on the Internet. Further examples are well discussed in [14].
Such a proposed framework for meta-services enables new possibilities to establish further non-commercial and commercial services. This situation may help to push the idea to become reality.
An overview of the proposed framework is shown in Figure 1. On the lower-left side, the unstructured content of the Web provided by content services (CS) is symbolized. Only browsing by following links and using local features provided by these services (e.g. site map, local search functionality, etc.) are possible at this level of the model. This reflects the current situation of the Web.
The next level describes the meta-services (MS), which may provide additional information of CS for users or applications. Examples of existing MS are robot-based search engines, search catalogs, discussion and annotation services for foreign content, rating-level servers, etc. It is quite obvious that the proposed classification is going to become blurred with daily usage, because one and the same service (or even content) can represent its own entities of information as well as meta-information related to other content.
As already discussed above, MS are mostly isolated. That means, users can hardly find proper (task specific) meta-information related to documents they are browsing. To counteract these shortcomings, the meta-meta-services (MMS), shown in the upper-left, are proposed and introduced. On the one hand, the task of this layer is to know about existing MS and their profiles (type of service, objectives, topics or geographical areas which are covered, target audience, etc.). On the other hand, MMS has to provide the proper information for the integration layer (see paragraph below). Therefore, MMS will provide substantial information and will enable co-operation and information interchange, which are not available at the time of writing. It is worth mentioning that MMS are also going to become blurred into other services. However, for an easy understanding this classification seems to be useful.
The integration layer (IL) manages the proper information requests as well as the processing, combination and visualization of the information from CS, MS and MMS. Furthermore, the IL also has to consider user individual needs in respect to their user profiles and the respective content.
Figure 1 also shows different clients on the right. A future-oriented framework has to include and manage a wide range of existing front-end applications, like common PCs, Laptops, PDAs, handheld phones, set-top boxes, Web terminals, voice browsers, etc. It appears to be obvious that the variety of front-ends stated so far requires and enables different technical solutions for the integration layer: front-end integration (e.g. browser plug-in, applet, using XUL, etc.), client-side proxy solution, server-side proxy solution and server application integration (e.g. server side includes, etc.). For meeting the nature of the Internet, consequently the IL has to be designed in a distributed manner.
As shown in Figure 1, communication between at least MS and MMS, between MMS themselves, between CS, MS, MMS and IL as well as between IL and front-end applications is required within the framework. Improvements can also be achieved if communication is possible between MS themselves as well as between MS and CS. For example, MS can inform each other that new or changed meta-information is available. The proposal for the communication layer is a multiple solution using sockets, socks and HTTP.
 |
| Figure 1: Overview of the proposed Framework |
Figure 1: Overview of the proposed Framework. The schematic graphic shows islands of unstructured content, existing and possible meta-services (e.g. annotations systems, external metadata server, etc.) as well as meta-meta-services (a kind of Yellow Pages service for meta-services, accounting services, user profiling and management services, etc.). The integration layer handles the interaction between users, as well as providing an integrated view of the Web content, information of meta-services and meta-meta-services. It enables the combination of content information and information from several meta services related to the users' needs.
The author of this article would like to emphasize that the valuable and novel point is that information from more than one MS can be managed and provided in combination with the corresponding content. The pre-selection of services related to topics and tasks by users are also taken into account. The combination of several MS as well as MMS can provide synergetic effects, which may empower users to select, navigate, manage and evaluate content on the Web.
The aim of this section is to discuss some imaginable improvements for pre-existing services and to give examples of novel services based on the possibilities provided by IFAEK.
Distributed metadata services could counteract the timid usage of metadata within document structures, which seems to be particularly useful for pre-existing content. It is conceivable that such services include human generated additional information, like a classification scheme, keywords, description, related information, author information, etc. Furthermore, automated or semi-automated systems can track content services and generate statistical information about creation time, modification time, modification history, link information to former and more recent documents, information about embedded object (e.g. graphics, sound files, etc.) and hyperlinks. The ideas discussed in this paragraph , for example, can be used to provide additional information to corresponding Web content rendered within or besides the content, enrich link information and can be combined with other meta-services. The later idea enables also a feature, which may enrich search results with additional information.
Existing services like subject catalogs, annotated link lists, quality-proved subject gateways, etc. can also be seen as special kinds of metadata. Consequently, they can be used for an enrichment of the content provided, may help users follow links of particular interest and can be used to classify the content. Useful information for similar documents can also be provided by these services. Based on the ideas of the services stated so far, a novel quality metadata service can be provided. The aim of this service is to help users get more relevant and quality information they are looking for. It is proposed to take into account descriptive metadata (e.g. information about the author, publisher, description, classification, etc) and evaluative metadata (e.g. diction of the language, target audience, etc.). The proposed service should be suited to the management of an entire Web server, areas of Web servers, single objects, and even parts of documents, whereby existing metadata of refined granularity overrides more general metadata. The provision of metadata should be a multi-step process. On the first level, authors are allowed to refer quality metadata to Web content. On the second level, librarians and subject domain experts check, edit, enrich and rate existing quality metadata records or enter new records of their own. Quality metadata can enrich corresponding content in the visualization process or allow the combination of a search request with quality metadata. Furthermore, authors related to the corresponding topics and to the target audience can be rated either by experts or by information extracted by content ratings.
Another interesting field of services can be offered by distributed annotations and discussion threads related to content, which also represent a special kind of metadata. The possibility of annotations and discussions of foreign documents provide a very democratic feature. Apart from that, such services can also provide descriptions and hyperlinks to foreign content. Therefore, for example, they may be used to build guided tours for particular communities. Of course, this basic feature is not novel, but the new aspect is that information from more than one service can be rendered related to specific tasks and proper demand. An interesting feature may also allow the definition of an information entity of foreign content, and follow changes or provide the history of the chunk of information. For example, a four year old research paper gives statistical information about the sea temperature, the meta-service provides timeliness data and the trend of the past few years. Another service can supply the users with related news information based on the topic of a document or even related to the context of an information entity. Users are able to select one or more services and define their particular needs. For example, a user wants to get annotations from the academic community related to computer science, but also likes to be informed in the field of e-commerce business by annotations from the CNN community at the same time.
A further interesting field of services are shared bookmark services. The idea of sharing bookmarks is not novel, but the combination of such services with metadata services may support users to manage their personal views of the Internet. On the other hand, users can classify their own bookmarks. These human efforts may be exploited to create a collective taxonomy. Furthermore, the shared bookmarks can be used by other services to provide similar documents.
It appears obvious that many combinations of the services discussed within this Chapter are conceivable. Numerous further services and combinations of such services are possible. It seems to be obvious that also computer processes of intelligent agents may exploit the provision of additional information by IFAEK.
On the one hand, a wide range of shortcomings can be identified in the broad field of retrieving and managing information on the Web. On the other hand, a lot of isolated research work and pre-existing non-commercial and commercial services providing additional information can be detected. This led the author of this article to propose IFAEK, a framework of meta-services and meta-meta-services, which allows the improvement of existing services and to build novel services. The key features are that users are enabled to define particular meta-services related to different subjects. Meta-meta-services manage the information interchange between meta-services and help users to define their favored meta-services. IFAEK enables the provision of additional information from more than one meta-service. Information may be provided within or besides the original documents as well as within the search process.
The aim of the framework is to reach a more structured and personalized Web. Some improved and new ideas are discussed. We have shown that the objectives of a smarter Web can be reached by means of a smart framework like IFAEK.
The most important step for future work is to reach the critical mass of research institutes, application developers and existing service providers. Together an open standard has to be defined and a proper mirco payment system for commercial services has to be established. The Web must be enriched with more structure and meta-knowledge to support users and automated programs. We have to make the Web smarter. The proposed system IFAEK could be a solution to help us reach the dream of a more structured, interactive and personalized Web, and may enable novel non-commercial and commercial services.
First of all, many thanks to Prof. Herman Maurer (IICM) and Prof. Frank Kappe (Hyperwave R&D), who have contributed many interesting aspects and a lot of useful hints. Also I would like to thank the members of the WAG-IICM for some valuable discussions. Many thanks also to Infodelio Information Systems and Guetl IT Research and Consulting for supporting this paper. Last but not least, many thanks to Mrs. Muml, Mr. Bukowski, Mr. Baer and Mr. Zweiigel for their remarkable support.
This article may be dedicated in memoriam to Mr. Franz Hesse, who had strongly influenced important parts of my life and my schooling.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Christian Guetl
Project Manager
Web Application Group, IICM
Graz University of Technology
Austria
CEO Infodelio Information Systems
Graz, Austria
Head of Guetl IT Research & Consulting
Graz, Austria
cguetl@acm.org
http://www2.iicm.edu/cguetl/
Phone: +43 316 876 5639
Fax: +43 316 876 5699
Christian Guetl is employed as a Project Manager at WAG-IICM
(Web
application group at IICM, Graz University of Technology). His
responsibilities include technical and organizational project
management in the field of information systems, establishing
internal project management and supervising of student's works.
He was co-initiator of the open source project xFIND
and
has conducted the lecture "Knowledge Management" at Graz
University of Technology for several years. Besides that,
Christian Guetl is head of Guetl
IT Research & Consulting
and CEO of Infodelio
Information Systems
.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
For citation purposes:
Guetl, C. "IFAEK: A Vision of Improvements for a More Structured and Personalized World Wide Web", Cultivate Interactive, issue
5, 1 October 2001
URL: <http://www.cultivate-int.org/issue5/ifaek/>
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Related articles:
If you would like to view similar articles to this one click on a key word below:
< - metadata - ifaek - personalized - Web - >
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
|
|