Meta-information for the World Wide Web

Jay Dhami

Department of Computer & Information Science
New Jersey Institute of Technology
University Heights
Newark, NJ 07102-1982

Faculty Advisor

Michael Bieber
New Jersey Center for Multimedia Research
Hypermedia Information Systems Research
New Jersey Institute of Technology
+1 973 596-2681 Fax: -5777


Currently there is no meta-information available for users on the World Wide Web. By providing meta-information for objects, users will have more information about the object. Users will also have information about anchors and links. Users could look at the meta-information about an object and decide if they want to visit that link or not. This research project provides a mechanism for displaying meta-information for objects on the World Wide Web and describes the importance of meta-information on the Web.


Meta-information is information about information. The use of meta-information can be seen very commonly in library databases. Title, author name, summary, publisher, etc. are examples of meta-information about a book. The project will use the same idea for the Web pages. By displaying such information, users will gain a deeper understanding of what they see as they navigate on the Web. Displaying meta-information will also help in reducing the disorientation of the user. The user will be able to find out what a document is about before even clicking on the link to take him or her there.

Currently there is almost no meta-information available on the World Wide Web. Most of the documents are being generated without meta-information. Meta-information existing within documents is inaccessible to the user as current Web browsers do not display it.

Annotation could also be part of a systemís meta-information. Since annotation is what the user writes for himself or others, it provides information about information. The user can note some key points, things that interested him/her, related work, etc. In contrast, most hypermedia systems provide for meta-information for the links and documents (nodes) [4]. Hypermedia data includes semantic node & link type, allow authors to add semantic type and annotations, and display these automatically to users [1]. Hypermedia research has been ongoing since the 1960ís.

The goal of this research is to produce a hypermedia engine uniquely designed to provide hypermedia functionality (e.g., annotation, meta-information, navigation, guided tours) to computational applications [8,9]. This will be done with minimal or no alternation of the application. This paper will try to describe the importance of meta-information in the World Wide Web and other hypermedia systems. The research also looks into different aspects of the meta-information. These aspects include dynamically generating meta-information, what to include in the meta-information, and different ways of displaying the information. The goal of this project is to convince developers of the next generation of browsers and systems to support more meta-information. Following sections describe the architecture of the hypermedia engine, the functionality of meta-information, generation of meta-information and global impact that the project is hoping to achieve.


Broad research goal of the project is to develop a hypermedia engine, which is called DHymE or the Dynamic Hypermedia Engine. DHymE runs in parallel with other applications and dynamically gives these applications automated linking and other hypermedia functionality. This other functionality includes annotation, guided tours, a history list, and the ability to display meta-information. The word dynamically means that there will be a run-time mapping of the application information and relationships to hypertext objects both within the system and external do it. An important feature of DHymE is to keep changes to the application itself to a minimum. The project proposes to incorporate this mapping with the help of logical rules called bridge laws. Bridge laws are used to create links and anchors from the elements extracted from an application (or Dynamically Mapped Information System or DMIS) by the DMIS Wrapper (DMISW).

Three aspects of our architecture are especially relevant to this paper. First, the project will identify application objects through the application's internal structure and not through lexical analysis (e.g., text analysis or keyword search). This enables us to get meta-information on each object from the application (DMIS or DMISW in Figure 1). Second, the project will display meta-information through our user interface wrapper, which means that the approach will work for any Web browser. Third, we provide meta-information on links as well as documents (Web pages). The bridge laws provide meta-information for both application objects and for links, which our user interface wrapper can package for display on the Web browser.

Many applications being brought to the Web have a well-defined architecture. A primary portion includes a computational section and a secondary portion includes the wrapper section which grabs the output of the computational part and formats it into HTML thus allowing it to be displayed on the browser. DHymE intercepts the messages (in this case, HTML pages) between the application's computational section and its interface section [8], automatically detecting the location of interrelationships in messages and documents based on the knowledge of the application's internal structure. Bridge Laws provide the mapping between the objects and relationships and hypermedia nodes and links. Every anchor displayed on the browser essentially leads to a set of all relationships from its underlying application object, as well as to any user-specified annotation and any other relationships external to the application.

One of the project goals is to be as much platform independent as possible. Therefore, Java was our language of choice due to its operating systems independence, networking availability, database connectivity, etc. Project also uses technology such as eXtensible Markup Language (XML). XML is designed so that systems using meta-information could take advantage of it.


Figure 1 describes the Dynamic Hypermedia Engine (DHymE) architecture.

Following are the explanation of terms used in the figure 1:

  • User Interface (UI): A World Wide Web browser.
  • User Interface Wrapper (UIW): The UIW has three main functions. First it translates messages from the DHymE's standard format to something the UI can process, and vice versa. Second, it passes messages between the two systems. It buffers the UI, implementing any functionality the engine requires of the UI, which the UI cannot provide itself. Third, it merges anchors and meta-information from the links and objects.
  • Gateway: All the messages that will be passed to different modules will have to go through the Gateway. Gateway will deliver messages for different modules.
  • Bridge Law Mapping Engine: This module will examine the list of elements in a document or screen generated by the DMIS and find links related to elements using bridge laws identified by the DMIS Wrapper.
  • Dynamically-Mapped Information System (DMIS): The DMIS is an application, which dynamically generates the contents of its displays, or otherwise requires the hypermedia engine to map hypermedia for it automatically at run time. An example of DMIS is one of the systems on which the project is being implemented onóan administrative Personnel Requisition Tracking (PRT) system at the New Jersey Institute of Technology.
  • Dynamically-Mapped Information System Wrapper (DMISW): The DMIS Wrapper translates messages from the DHymEís standard format to something the DMISW can process, and vice versa. It also passes messages between the two systems. If necessary, it identifies and marks elements in the DMIS messages to which hypermedia components are mapped.
  • DMIS Instances: These are the individual programs, worksheets or other application instances that run under the DMIS system. For example, an expert system shell or relational database management system (RDBMS) would be a DMIS whereas the expert systems written in that shell or the databases manipulated by the RDBMS would be DMIS instances. In our case the PRT system is a single DMIS with no instances.


All the modules will generate meta-information. Meta-information about documents, anchors, and links will be generated by the DMISW. Some of the meta-information might be object type, link type, link label, object label, etc. Documents generated by the DMIS will be processed by the DMISW, which would identify possible elements for linking with Element Identification Rules. Element Identification Rules are rules which help in figuring out which DMIS instance elements appear in the DMIS generated documents. The bridge law mapping engine will use Bridge Laws to identify links for the identified elements. There are two kinds of Bridge Laws: Link Bridge laws and Regeneration Bridge laws. Link bridge law will map element type to links. Regeneration bridge laws will regenerate an element according to the element ID and its meta-information. Meta-information about anchors will be stored in elements and meta-information about links will be stored in Bridge Laws. Meta-information will be embedded with the anchor information.

Once meta-information regarding an object is generated it will be stored in a database and displayed to the user. Some of the information, which might be generated dynamically, is the information about bridge laws, relationship information, and other information which might change. Additional information would be the date on which the document was created, last modified, and other information about the document. Meta-information that would need to be inserted manually includes author name, description information about anchors and links, and any other additional information useful to the user or author could be included. There are also ways to automatically generate a summary of the document, which would be regarded as part of meta-information.

The meta-information will be stored in a database in the Engine. The database will contain all the meta-information about all the DMIS. Whenever the meta-information is needed it could be called from the database through a database call within our common message format.


There are many ways to display meta-information in Web browsers. Meta-information can be displaying using pop-up windows, which would display meta-information when a user moves the mouse over a anchor/link and the window will disappear when user moves mouse away from anchor/link. The project will use a three-frame structure which would contain all the information. The first frame would contain the main document. The second frame will contain list of links related to the anchor. The third frame will contain meta-information about objects.

The project will display meta-information to a frame when a user moves mouse over an object. To accomplish this, project is using JavaScript. JavaScript has event handlers which will enable us to do anything when the event occurs. We are using evens onMouseOver and onMouseOut. OnMouseOver is an event handler which will "watch out" for when a user moves his mouse over an anchor. OnMouseOut will "watch out" for when the userís mouse leaves the object. Following is the script which will display the meta-information held in the parameter "message" to the frame.

function DISPLAY_METAINFO(message)
var framed = parent.FRAME3.document;;
framed.write("<BODY BGCOLOR='WHITE'>"+"\n"+"</BODY>"+"\n");
framed.write("Meta-information about the link/Anchor".bold().fontcolor("BLUE")+"\n");
framed.write("<HR>" + "\n");

Figure 2: Sample code for displaying meta-information

This function of JavaScript will write the meta-information to its designated frame.


One feature that could be added is to allow users to add meta-information, instead of having owning applications generate it. This could also include annotation. This additional meta-information could be stored two ways. One of the ways is to store it in the meta-information database. The new meta-information could be added to the object's initial information. The other way to store meta-information is in the user's local database. The user could include anything he/she may want to include. This also includes annotation that might help user with reading of the document information.

Adding additional information would also bring up an issue of security. There are number of reasons for adding access control. Some documents might have confidential meta-information which might not be suitable to display it to everybody. Access rights would able to distinguish the permissions on a user. Also, who should be allowed to add meta-information? There should be some kind of access rights on who could add more meta-information and who may view it.

Since users will be given with considerable amount of meta-information, there should be a userís preference module. This module will keep track of userís preferences. A user can take advantage of user preferences in displaying of meta-information. A user should be able to specify how he/she wants to view the information. For example, If a user wishes to view only authorís name and document title, then the user could set up his/her preferences and information will be displayed accordingly.


Current Web browsers do not automatically display meta-information. The new generations of browsers will need to be able to do this. The new browsers should have a mechanism for displaying, adding, and user preferences about the meta-information. Therefore there needs to be standard form of encoding meta-information so that browsers can automatically display anything encoded in this way.

Since most of the meta-information might need to be added manually by the authors, there should be an authoring tool, which might help the authors add meta-information to their documents. This authoring tool could be added in the next generation of browsers. This standard feature in the browsers will make it easier to enter meta-information about an object. Other applications will be able to automatically extract meta-information. By having standard routines for helping application builders tell browsers where to find this information.


As the use of the Internet continues to grow, and other systems which involve massive amount of information, hopefully there will be more systems which display meta-information. Furthermore, hopefully meta-information will become a standard feature of browsers.


1. Bieber, M., Vitali, F., Ashman, H., Balasubramanian, V., and Oinas-Kukkonen, H.: Fourth Generation Hypermedia: some missing links for the World Wide Web. International Journal of HCI. 1997.

2. Ramanathan, P., Pattabiraman, M., Lu, Q., and Bieber, M.: Applying Hypermedia To World Wide Web Applications. Technical Report. 1997.

3. Ramanathan, P., Bieber, M., and Rana, A.: Providing Meta-information for World Wide Web Links and Documents. AIS'97 Conference.

4. Conklin, J.: Hypertext: An Introduction and Survey. IEEE Computer. 1987.

5. Lassila, O: Introduction to RDF Meta-information.

6. Weibel, S., and Miller, E.: Dublin Core Meta-information Element Set:

Resource Page.

7. Whitehead, E.: An architectural model for application integration in open hypermedia environments. Hypertext 97 Proceedings, ACM Press, New York, 1997, 1-12.

8. Bieber, M., and Kacmar, C.: Designing hypertext support for computational applications. Communications of the ACM 38(8), 1995, 99-107.

9. Bieber, M., Supplementing Applications with Hypermedia, Technical Report, CIS Department, New Jersey Institute of Technology, 1997.

10. Davis, H., Hall W., Heath I., Hill, G., and Wilkins, R; Towards an integrated information environment with open hypermedia systems. Proceedings of the ACM Conference on Hypertext (Milan, Nov. 1992) 181-190.

11. Hill, G. AND Hall, W.: Extending the Microcosm model to a distributed environment. European Conference on Hypermedia Technologies'94 Proceedings (Edinburgh, Scotland; September 1994), ACM Press, 32-40.