By choosing to view this document, you agree to all provisions of the copyright laws protecting it.
|Peter W. Flur||John B. Lockhart||Sudhakar Yalamanchili|
|(404) 89404767||(404) 894-2925||(404) 894-2940|
|Georgia Institute of Technology|
|School of Electrical and Computer Engineering|
|Atlanta, GA 30332-0250|
Recent developments within the World Wide Web (WWW) and attendant user interfaces (e.g. Mosaic, Netscape) have produced well-defined protocols for describing, communicating and viewing hypertext information. This ability to uniformly handle different types of information has created tremendous opportunities for re-engineering the means by which disparate information is managed and communicated among individuals and organizations. In the School of Electrical and Computer Engineering (ECE) at the Georgia Institute of Technology, this technology is being used as the building block for integrating administrative, instructional and research services. This paper describes tools, applications, experiences and observations surrounding the development of the ECE Web Server.
The World Wide Web was originally conceived at CERN (Conseil Européen pour la Recherche Nucléaire, http://www.cern.ch/) in Switzerland in the late 1980's as a method of distributing hypertext documents internally among physicists at the Particle Physics Laboratory. The phenomenal growth in the World Wide Web outside of CERN started when a few undergraduate students at the University of Illinois (Urbana-Champaign) National Center for Supercomputing Applications (http://www.ncsa.uiuc.edu/) created a Web browser called Mosaic. Mosaic was designed to use the protocols and specifications developed at CERN to enable users on all platforms to exchange, access, view and manipulate distinct types of information via a computer network using a simple point-and-click interface. Since Mosaic's first release in early 1993, both the number of World Wide Web servers and the volume of World Wide Web traffic on the Internet have grown exponentially.
The ability to handle different types of information has created
tremendous opportunities for reengineering the means by which
information is managed and communicated among individuals and orga-
nizations. Like many other organizations, Georgia Tech (http://www.gatech.edu/) has
experienced explosive growth in the provision of WWW-based access by
Schools and individuals to distinct types of information. However,
perhaps unlike many efforts, the School of Electrical and Computer
Engineering (ECE, http://www.ece.gatech.edu/) at
the Georgia Institute of
Technology is utilizing this technology as the building block for
integrating access to administrative, instructional and research
services. This paper documents not only some of the tools we have
developed, but also our experiences, lessons and suggestions.
Traditionally, academic, administrative and research services were available through distinct channels and as a result duplicated substantial amount of information, requiring significant coordination to maintain consistency. Evolution of distinct offices for services is partly historical, as well as motivated by the distinct needs of each service. For example, security requirements for financial and personnel information are distinct from that of curriculum information. Our WWW services integrate access to information used in all three services. This enables the School to present timely and useful information to its students in a consistent, efficient, intelligent manner, and to provide a common base of knowledge that can be shared by alumni, prospective students, research sponsors, educators and administrators. This technology provides new avenues to disseminate information and thus impact not only traditional class- room instruction, but the complete educational information infrastructure.
|Figure 1: Unified Interface to Campus Information|
The ECE Web (see Figure 2) server came on-line on an exploratory basis during the Spring of 1994. Early uses of the Web were administrative and general in nature, providing ways of disseminating routine information such as phone and electronic mailing lists. The first major step came in 1994 when we began integrating our extensive departmental relational database with the Web server. This provided a wealth of information about all facets of the School for many different groups of people. We focused on administrative and academic applications during the first six months. Since that time, we have developed a suite of applications which are actively used in all areas of the School beginning with the student and faculty recruitment processes, including academic affairs (such as course descriptions, outlines, syllabi, grading scales and even lectures and homework assignments), administrative and financial services, as well as sponsored research.
|Figure 2: ECE Home Page|
Academic services on the ECE Web server include registration-oriented applications, information relating to specific classes as well as general academic information. Among the registration-oriented applications included on the server are course listings, general outlines, information on course scheduling and links to faculty-related information. Faculty have access to class rolls and related academic information. This information is maintained in a relational database that is protected via authenticated access (discussed in greater detail under Administrative Services).
Course listings show all courses offered by the School, with links to course abstracts taken from the General Catalog and detailed outlines taken from the ECE Student Handbook. Detailed outlines include a systematic breakdown of all material covered in the course, links to prerequisite courses, textbook information and links to course scheduling information. Course prerequisites may be viewed textually or graphically, as shown in Figure 3 (http://www.ece.gatech.edu/cgi-bin/prereqs.cgi?ee6414), with either method allowing the student to dynamically access information on any of the prerequisites. Course listings provide scheduling information dynamically generated from the campus registration system. This information includes the number of students registered and the maxi- mum permitted class size. Links to relevant information such as the faculty members's home page are also provided. This tight coupling with the campus registration system ensures that students have current, up-to-date information, and precludes constant cross-checking among individuals and organizations. The normal paper delays in the migration of up-to-date information from the campus records to departmental records are eliminated.
|Figure 3: Hypertext Graphical Course Pre-requisite Tree|
The links to faculty home pages are helpful because they not only include photographs, personal biographies and links to research pages, but they also provide prospective students with the ability to look at past and future offerings among the various faculty. This information could be used to determine professor's work loads in any given quarter.
All course-related information is provided in hypertext form. This allows the student to travel from a list of classes offered in one quarter to the outline of a particular course the student may be interested in. From the course outline, the student may view course prerequisites, examine a prerequisite course's outline, and determine when the prerequisite is offered. The student might visit the home page of the professor teaching the course, or examine the textbooks that are required for the course. The School's Web interface to course information provides the student with a powerful tool to quickly make informed choices when planning a course of study. As the School's course information is globally and publicly available, it also serves to inform prospective students of the strength and depth of the curriculum.
One of the more exciting applications of the Web has been the ability to generate and provide a new form of on-line tutorial instruction. For example, EE 2200, Introduction to Signals and Systems (http://www.ece.gatech.edu/research/DSP/courses/ee2200/), is a recently introduced first course for Computer Engineering majors. The Web is used to provide tutorial instruction, laboratory assignments, homework solutions and lecture notes. Students can view waveforms, listen to the application of classroom material to audio signals and view the results of elementary signal processing operations in real-time. By combining textbook and traditional classroom instruction with immediate feedback from real applications, student learning is significantly enhanced. Making these facilities available from their personal computers in their dormitory rooms precludes any need for expensive laboratory facilities and their maintenance. The academic Web pages have been structured and customized to encourage faculty to continue to develop such Web-based instructional material where beneficial.
Although Web-based course information primarily serves on-campus students, much of the information is also externally accessible and can be used to audit courses. Many of the School's courses are videotaped in support of remote instruction and distance learning. When the Web is used to promote course information and remote students have Web access, the administrative overhead of teaching a video section is dramatically reduced. Video students, previously encumbered with large mailing delays, suddenly can become more active participants in the course.
In addition to registration and course-specific information, the School Web server functions as a general bulletin-board for other School-related academic information, such as regulations, policies, laboratory hours, proposed or projected course offerings, committees, etc. Student groups such as the IEEE and ECE-related honor societies, e.g., HKN, also have pages on the School Web server. The server includes graphical floor maps of buildings associated with the School, with individual rooms linked to information about the room. Room information includes links to the home pages of personnel in an office, or in the case of a public laboratory, lab information, schedules and sponsor information. Telnet links are provided to individual machines in the lab and links to on-line application forms for computer accounts eliminate traditional account forms. All computer accounts are requested, authorized and created on-line.
Since the early 1990's, the School of Electrical and Computer Engineering has undertaken a major software development effort to utilize the capabilities of a powerful relational database (Oracle) to help offset a rapid growth in faculty size without a commensurate growth in support staff. Since 1987, the School of ECE has added more than 30 faculty members with a corresponding substantial increase in sponsored project funding. Prior to 1992, all accounting was done via standard Georgia Tech business office reports and manual ledgers kept within the School. With more than four hundred active projects in the School, this quickly became unmanageable.
Many options for a user interface to this database were available, but Oracle Forms 3.0 was the platform originally chosen. Although an extremely valuable development tool, Forms is quite complex, and it was unreasonable to expect general faculty and staff to become proficient in its use without extensive training and individual support. When the potential of the Web became evident, it was obvious that it would become the access tool of choice for gaining financial information and reports from the relational database. Our faculty now have real-time access to all financial data for their sponsored research projects through the ECE Web server. We have built the necessary security tools to protect the information from both external intruders as well as unauthorized access from internal faculty, staff or students.
The Web-based services include a variety of project tracking and financial reports. Faculty and staff are presented with a menu of report choices, then asked to specify configuration options for specific reports. These options might include control over sections of the report to be generated, a time frame to limit displayed transactions, a list of projects for which a report may generated, etc. Listings of projects or employees are limited by the access level of the user of the system. For example, while a member of the accounting staff may generate a report for any project, a secretary may only generate reports for projects corresponding to faculty or staff to which the secretary is assigned.
Reports include everything from exhaustive project summaries (including budgets, balances, personal service charges, transactions and audit information) to more specific reports such as personnel information, document (purchase order and travel request authorization) information, etc. Hypertext links interconnect virtually every report. For example, from within a project transaction summary, a faculty member can view a report on the specific purchase order or travel authorization associated with a particular charge. These reports include actual graphical representations (images) of the purchase order or travel authorization form requested (see Figure 4). Even starting with a report on a specific purchase order, the faculty member can view the summary report for the project to which the purchase order is charged.
|Figure 4: Purchase Order and Travel Request Authorization Form Images|
Tracking graduate students and faculty participating in large research projects is now a few keystrokes away. Faculty may view summaries of their own charges, or use links within those summaries to view reports on the projects to which the charges apply, depending upon their access privileges. Faculty viewing their own projects may see summaries of graduate student charges related to the projects. Wherever possible, charges corresponding to personnel are linked to the home page of the person in question. One significant achievement has been the ability to reconcile the local books with those of the Institute in real-time and on demand. The faculty can view reports from each perspective and then select automatic reconciliation which will examine each report for discrepancies. This has dramatically reduced the amount of time wasted in manual reconciliations.
Security is a major concern with any application developed for use on the Internet, especially with applications including sensitive financial information. Without the necessary resources available to purchase commercial a World Wide Web server which could easily encrypt client/server transactions, we were forced to build our own security measures. Using public domain encryption utilities and libraries, we retrofit public domain World Wide Web servers to provide secure user-based authentication for our security-sensitive applications and reports. We have also investigated commercially available secure servers as a means of providing even greater security.
Despite the internal complexity of the database and access protocols, the rich graphical format of reports and availability of system security, the Web-based approach allows a consistent, uniform interface across many types of computers throughout the School's distributed network. The Web's consistency and ease of use dramatically reduced training time and indeed was the major factor in faculty acceptance of the database system. Personal Computers, Macintoshes and several varieties of UNIX Workstations all operate with a consistent interface to Academic and Administrative services.
Dissemination of academic research results is one of the first and most widely used applications of Web services. We have provided a consistent organizational structure for research information within the School's Web pages which gives control of information content to the research groups and individuals themselves, while still allowing easy information retrieval. This information is provided on our server (http://www.ece.gatech.edu/research/) arranged hierarchically by domain (e.g., Computer Engineering, Telecommunications, etc.). Each domain has an interface consistent with the common layout as well as with the School's printed research publications. A Web user can easily access general, academic (courses of study, class information, etc.) and professional information in any of the technical areas.
Given control of information content, many research groups have responded by providing Web summaries of current research projects, often including complete sets of related publications and abstracts on the Web. While in the past faculty and graduate students provided their own publications and project summaries on their individual home pages, these publications are now also linked via research group pages, allowing straightforward retrieval by the Web user. Even files which were available solely on the School's anonymous FTP (File Transfer Protocol) server are now linked from the School's research Web pages.
The School's research pages are quickly becoming a source of
information for potential students and new faculty. The pages are now
the primary source of information for generating other documents such
as research summaries, domain-specific technical brochures and
sections of the faculty handbook.
What is hidden in much of the above description is a significant amount of effort expended towards the development of tools and interfaces. We have built a variety of tools that provide much of the automation currently available within our School servers without requiring extensive databases.
One example is a class of tools we have developed to process dynamic image maps. Image maps were first introduced to the WWW as a means of providing a graphical interface to static information. The first image maps on the Web required several steps to implement: preparation of the image in a graphics utility, specification of active map regions on the image via a image map editing tool, then installation of the image onto the server using yet another tool to process image map information. Should anything need to be changed, the whole process would have to be performed again. This entailed significant manual effort both in terms of time and computing resources.
ECE has developed several tools around a single utility which
automates this three step process. The utility reads a control file,
dynamically generates an image from the control file on demand, deter-
mine coordinate pairs and bounding boxes, and computes Web locations
corresponding to image map regions. We are using this utility in at
least three different parts of our server, including the button bars
that are prevalent throughout our pages, the interactive maps which
allow users to see a graphical representation of a building and
click on rooms to see where people are located, and finally, the
interactive prerequisite tree generator described above.
Although we initially encountered many stumbling blocks to the acceptance of the Web (notably the expected early resistance to anything new!), Figure 5 shows that the information content available on the Web server is valuable to parties both internal and external to our School. Usage of our server continues to dramatically increase, both by information seekers and providers. It is evident that the Web provides an excellent vehicle for organizing and presenting vast quantities of dynamic information as well as encouraging totally new applications.
|Figure 5: ECE Web Usage|
We have learned many lessons in the development and maintenance of our WWW server. First and foremost in an era of decreasing budgets, automation is a key to success. The key to automation in this environment remains a detailed relational database which is normalized and extremely broad in its information content. Lists such as rosters and phone lists are always out of date the minute they are created unless there is a central authority that maintains that information in the database. Financial and curriculum information is in a constant state of change. We have been fortunate to be able to grow the back- end database and distribute the authority in all cases to the custodian or originator of the information. This allows the owner to maintain the data and be responsible for its correctness. As Webmasters, we prefer to consider ourselves as maintainers of information repositories rather than information providers. We simply provide the means for making available information that is critical to both the School's operation as well as its long term success. We allow custodians of the data to present in a useful and timely manner to new customers all over the world without regard for national or international borders.