This paper is the intellectual property of the author(s). It was presented at EDUCAUSE '99, an EDUCAUSE conference, and is part of that conference's online proceedings. See http://www.educause.edu/copyright.html for additional copyright information.

Lynn M. Lickteig
Photographer V and Honorarium Instructor
College of Architecture and Planning
University of Colorado at Denver
Campus Box 126
P.O. Box 173364
Denver, CO 80217-3364
[email protected]

Robynn F. Tripp
Information Technology Professional II
Information Technology Services/Applications Information Services
University of Colorado at Boulder
Campus Box 455
Boulder, CO 80309-0455
[email protected]

Abstract

Faced with the challenge of providing access to an architecture slide collection to students and faculty on two campuses, the University of Colorado developed an application utilizing state-of-the-art client-server and Web technology to catalog and display information on over 27,000 slides. A collaborative effort of the College of Architecture and Planning and the Information and Technology Services Department, the Visual Information System of Colorado (VISC) provides information about the slide subject and circulation and reservation information about each slide. A lecture module included in VISC enhances faculty support by enabling faculty to reserve and assign slides for specific lectures. This paper will provide an overview of VISC, a summary of the principles underlying its success, and lessons learned during development of the project.

University of Colorado

The University of Colorado is made up of Central Administration and four campuses, including the University of Colorado at Denver, the University of Colorado at Boulder, the University of Colorado at Colorado Springs, and the University of Colorado Health Sciences Center. The Boulder campus, founded in 1876 offers more than 150 fields of study across ten colleges and schools. The campus community includes 2,982 staff and 1,175 faculty members in 270 departments supporting over 25,000 students. The Denver campus, founded in 1912, offers approximately 88 fields of study across eight colleges and schools. The Denver campus includes 732 staff and 763 faculty members supporting over 11,000 students.

College of Architecture and Planning/Visual Resource Center

FIVE TRUTHS: THE CLIENT�S PERSPECTIVE by Lynn Lickteig

I have served for 15 years as Director of the Visual Resource Center at the University of Colorado at Denver�s College of Architecture and Planning. Our graduate programs are located in Denver, Colorado, and our undergraduate program is delivered on the Boulder campus. We created a computerized text database for our undergraduate slide collection in 1988. Through a modest $13,000 grant, we paid a private contractor to design this database for us in dBase III. As we expected, the software became outdated, and we wanted to upgrade it to handle image display as well. This pilot database fueled the idea that the College of Architecture and Planning should develop its own image database. We had the cavalier thought that creating another database ourselves would be easy, since we had already worked with a programmer to create the first one. We went looking for funding, and in 1996 received a grant of $35,000.00 from the University of Colorado President�s Initiative Fund. My college teamed up with CU/Boulder's Information Technology Services department, which provided matching funds. We then spent two years together creating an image database using Oracle software. We named our product the Visual Information System of Colorado, and refer to it using the acronym, "VISC." After VISC was developed, a second grant provided money to develop a Web interface that we call the Electronic Library of Colorado Architecture, Landscape, and Planning (http://www.cudenver.edu/public/AandP/ELCALP/). The Electronic Library pulls its data and images from the underlying VISC database, and is a good example of how one technology project and one grant can lead quite naturally to another.

Serving as the project manager for the development of two databases and a web interface in the last ten years has led me to realize some simple truths about the factors that must be present for an educational technology project to be successful. From the outset, I want to state that while these truths are extremely simple in theory, many institutions of higher education, particularly in the humanities departments, do not embrace them as fully as they should.

The first truth about technology projects like VISC and our Electronic Library Web site is that you can not do it alone. Collaboration is a must. You will need a small team of people, including everyone from the administrators who control the budget and give the project their political blessing, to the IT programmers who write the code. Also essential are the staff members who enter the data and scan the images, and the faculty members who provide professional input and write funding grants. The collaborative team approach seems so logical, but time and again my colleagues in the Visual Resources Association tell stories of administrators who assume a staff member should be able to single-handedly pull off a technology project, with no decrease in other duties, and no additional staffing to help carry the daily load. And faculty members may voice a willingness to help at the discussion stage, but since technology "service" projects often count very little towards their tenure, promotion, or salary reviews, most faculty members are not willing to give the two to ten hours per week or per month that a technology project might demand from them. And can you blame them? Their success lies in research, not service projects. Clearly the system of rewards would have to change before humanities faculty members embrace serving as mentors for educational technology projects.

The team of people it took to develop the VISC database and the Electronic Library application in the past three years has included: two faculty members, two classified staff members, two technical support staff, five undergraduate students, three graduate students, one Ph.D. student, the five employees of ITS who wrote the VISC database, one private consultant who designed the Web site, one digital image consultant who taught us to scan and set up the equipment, and finally, one content consultant who wrote narratives on buildings and architects, and who helped us catalogue new slides. And for those of you who wonder what the Full Time Equivalent adds up to for all these individuals, I�d estimate about 7.5 FTE from the university, not including the three private sector consultants. And it needs to be said that collaboration involves more than one or two token meetings between the IT staff and the client. It took two years of bimonthly, one-on-one meetings to hash out and test the functionality of the system and the layout of the screens. I want to emphasize that the VISC was a project that we all took on in addition to performing our normal job duties.

The second truth is that technology projects always require more money than you think they will, at least more than most of us in the arts or humanities fields of academia have in our operating budgets. Additionally, it seems that funding for equipment and software is always more readily available than funding for staff wages. My advice is to calculate what you think you�ll need and add at least 50%. So far, we've been fortunate to receive over $120,000 in technology grants during the last ten years to develop the VISC database and the Electronic Library Web site. Yet, I think the primary mistake that higher education institutions make is the dependence on grant money to fund technology projects. Many times I have been told, "great idea, now go out and get a grant--our general fund can�t support it." In truth, my college was also guilty of this mistake.

A source of tension between us and our ITS collaborators was that most of the money for the VISC project came from external grant funding. We had no college general fund money earmarked to pay ITS for system enhancements, cost overruns, or yearly maintenance. In fact, we naively thought that since CU�s ITS staff members had created the database--and we�re all one family--that they would surely maintain it free of charge! This was not their understanding, however. Furthermore, grant givers imposed fixed expenditure categories, including stipulations regarding who could be hired to work on the project. What I�d like to advocate quite emphatically is that administrators should treat technology projects with the same seriousness and attention they would give academic accreditation reviews. They would earnestly correct deficiencies by engaging in fund raising or diverting money from other sources to pay for additional faculty, to make curriculum changes, or to correct safety code violations in their buildings. They should also earmark permanent funding sources for technology development as part of their yearly general fund budget requests. The message should be taken to heart that technology development is not a finite process and a matter of a one-time fix. One grant isn�t going to be enough to do the job in the long run.

A third truth is that calculating the amount of staffing needed for an image database project is always grossly underestimated, particularly if the faculty members and staff members involved are expected to carry out their "other" duties while engaging in technology development. We estimate that one slide can be scanned in a few minutes. Mounting the slide, applying and printing the labels takes another few minutes. Data entry takes ten more minutes. However, I advise you not to naively calculate the wages for staffing based only on these criteria. Do not forget to include money for staff members to catalogue the images prior to data entry. Do not overlook the simple fact that the text data associated with the image doesn't jump into the computer with the same ease as the scan, and that research must be done to verify even the simplest data.

The fundamental fact is that the importance and integrity of an educational image database lies in the text database first and foremost. Students and faculty members require data of great integrity. Other institutions and our clients will judge us based on the quality of our information. In an educational image database, a picture is not worth a thousand words. But according to stories I have heard from other slide curators, many administrators put the cart before the horse and want their staff to do imaging projects without a database to manage the images. You wouldn�t buy 10,000 books and store them in a warehouse, and call it a library. Similarly, to embark on an image scanning project without a database to manage and retrieve the images, and without scholarly text to educate your clients is foolhardy to say the least. Yet I heard one administrator say, "I don�t care so much about the cataloguing or the database, just give me images." Another asked me, "can�t you get a student to write you a little program?" And another: "Maybe you could take a week and learn a software program and write your own database." Somehow these administrators thought that simply having pictures on a computer screen meant that their departments were technically savvy. To them, the difficulty of creating a database, as well as the importance of the words accompanying the images, were lost.

A fourth truth that I learned is that there is a language differential between Information Technology personnel and their non-technical clients. Even with our previous database as a model, conveying to IT personnel what we wanted the final VISC product to be was not always easy. One example occurred when I was discussing with one of the ITS programmers the various reports we wanted our database to perform for us. I said we needed simple reports, such as how many images circulated per year, which slides were used most often and by whom, and so on. When the programmer met with me to demonstrate how the reports would function, I asked the question, "Now, how do I print them?" He stared at me for a moment as if I was daft and replied, "You asked for reports, you didn�t say you wanted them to print out." The word "report" did not imply the same functionality to him as it did to me. Thankfully, there were few instances like this. Yet this example underscores the need for both sides to communicate clearly, to reiterate what is expected time and again, and never to assume that words hold the same meaning for everyone.

The fifth truth I learned is that there is no such thing as "the perfect" equipment. It's going to be out-dated in two to five years any way, so go ahead and pick a brand, based on advice from people you trust and your institution�s guidelines. But even then, cosmic rays will intervene so that of two identical new computers, one will run smoothly and the other will not. So expect equipment and software malfunctions, and don't browbeat yourself when you later decide you chose the wrong brand. And, of course, treat the IT and technical support staff with great respect.

Another area that people get hung up on is whether to use a PC or a Mac, and what software to use. The reality is that these decisions are probably going to be decided by someone other than the project director, for a variety of unrelated reasons. So, I�d advise you to give up any personal attachments you may have, and go with whichever platforms the people who hold the purse strings will support. In our case it was PC and Oracle. Though, truthfully, our project has involved both PCs and Macs for different tasks. The prototype databases were written in dBase III and FileMaker Pro.

My final piece of advice is that you can�t let the magnitude of the technology project, whether it be the staff time involved or the costs, keep you from beginning the journey. Our reputations as institutions of higher learning are defined not only by the caliber of our faculty and curricula, but also by the sophistication of our information technology resources for instruction and research. Students and faculty members increasingly expect to access educational resources online. Not to have our image collections and libraries computer accessible is seen as substandard. We must begin the journey, even if the first steps are small.

We are fortunate that the President of the University of Colorado, Dr. John Buechner, understands the importance of technology development in every department and college on our four-campus system. He is leading an initiative, called the Total Learning Environment, which includes among its goals, furthering technology projects in the academic environment. The TLE program embraces many of the same truths that I have mentioned, including the aspect of faculty mentoring. At a June 1999, budget retreat, President Buechner outlined several strategic TLE priorities, one of which emphasizes: "Maximizing faculty contributions, with an emphasis on redefining and differentiating faculty roles, clarifying contributions through service; integrating teaching, research, and service responsibilities; and an enhancement of the reward system." Another TLE priority is: "Entrepreneurial program delivery, which will target specific markets, commercialize learning technology, expand the use of enterprise models, encourage business approaches, increase revenue streams and maximize partnership opportunities." (CU Silver and Gold Record, June 10, 1999.)

We believe that the VISC database and the Electronic Library Web site are examples of the TLE initiative in action, and evidence that the University of Colorado is embracing the opportunity to move educational resources into the digital and perhaps commercial realms.

Information Technology Services/Application and Information Systems

An IT perspective by Robynn Tripp

Technical Specifications

The system server for VISC is run on a Windows NT 4.0 platform. Data is stored on the server in

an Oracle 7.3 database and client side user screens were developed in Oracle Forms 4.5. Reports and adhoc queries were built using Oracle Reports 2.5 and Andyne’s GQL. Image files, currently being created in Adobe PhotoShop 4.0 in Jpeg format, are stored on the server in a separate file system in three different jpeg resolutions. Presenting data with an actual image on the client side application involves accessing the data from the database and integrating it with the image file from the server file system. Security on the system is controlled through the client-side application, the NT operating system, and the Oracle database. The minimum client configuration for VISC users is Windows 95 or NT and at least 16 MB of RAM.

The ELCALP web site was built on a Windows NT platform using Microsoft’s Internet Information Server (IIS) and an Oracle database server. Customer Active Server Page (ASP) programs handle search requests and perform queries on the VISC Oracle database. Views, or subsets, of the Oracle database were constructed specifically for the ELCALP web site to facilitate faster retrieval on the information. Results of those queries are parsed, formatted, and returned as HTML to a web browser, enabling the user to browse the text and images housed in the electronic library.

Development of the system

The core application development team was comprised of two project managers from Information Technology Services (ITS), a project manager from the Visual Resources Center with slide library curator experience, an Oracle Database Administrator (DBA) from ITS, two programmers from ITS with GUI development experience, another slide curator from VRC, and two faculty members from the College of Architecture and Planning. Additional staff from ITS and the College of Architecture and Planning were pulled into the project as needed. The goal of the development team was to come up with an application that would be a logical replacement for their current system, able to use the converted data from the current system, user friendly, easily understood with flexible entry and query capabilities, and responsive to changing technologies. The sheer number of slides and the information that needed to be gathered and maintained on those slides dictated that the data be kept in a robust database that ran on a stable platform. Oracle was chosen not only because it could handle the complex and large amount of data expected, but also because Oracle was an emerging standard for the University.

Once the development environment was established, the data model for the system had to be constructed. Several months were devoted to this task, which involved defining table structures to hold the data, relationships between the tables, and joins and constraints on the table columns. Screen prototyping began after the data model was completed. Initially, the field layouts on the screens were sketched on paper and reviewed by the core group. Once a rough design was agreed upon, a prototype of the screen was developed with the Forms development tool. Members of the core group worked with the prototype screens to familiarize themselves with the "look and feel" of the GUI environment and to make recommendations on navigation wit hin the application and functionality as a whole. Development of the screens also became a testing point for the data model and there were times when the data model was revised to streamline or provide additional functionality to the application. Data from the previous system had to be extracted, converted when necessary, and loaded into the Oracle tables. Upon completion of the screen development and data conversion, the screens were packaged together in specific modules for cataloging, circulation, and system maintenance. Reports and queries were written and integrated with the modules where applicable.

VISC was deployed to the Visual Resources Center on the Boulder campus in the fall of 1997. The primary users of the system had been involved as core members of the development team and were able to help ensure the success of the implementation. There were modifications and additions that were made to the system by ITS, as well as data clean-up from the previous system that had to be completed over the next several months and in the spring of 1998, the application was implemented on the Denver campus.

Key points regarding development and implementation

When the project was undertaken and the core development group was formed, the goal of the project was supported not only by each team member, but by the Dean and Associate Dean of the College of Architecture and Planning, several faculty members from the College, and top level managers within Information Technology Services. Because there was a common, shared goal that was widely supported, there was a collaborative, cooperative environment generated for the development of the system. Each member of the core development group had their own distinctive perspective on the project, yet there was an optimum amount of interaction, sharing of ideas, and an understanding that there was a common goal behind the work. There was also an underlying idea beneath the project that the application could later be used by other similar departments on campus, such as the department of Fine Arts, for their slide libraries.

The actual application development process to create the screens and write the code to provide the functionality took approximately 5 months. The two programmers involved in writing the code had some experience with GUI programming, but had never worked with the Forms development tool and some of the development time was given to overcoming the learning curve on the new tool. Progress on development was sidelined when the data model had to be modified slightly to improve functionality of the application.

Funding of the project was also an issue during development. Initial funding for the project came from a grant secured by faculty and staff members from the College of Architecture and Planning and was

matched by funding from Information Technology Services. The initial funding was still not enough to cover the entire cost of developing the application and, again, momentum on the project was slowed as the members of the development group scrambled to get additional funding to complete the project. Funding for maintenance, continued development, and revisions for VISC remains a problem even today. Grants cannot be obtained for these tasks, general funding within the University is not available, and the current ITS budget cannot accommodate the amount needed on a year to year basis.

Finally, implementation of the system on two campuses proved to be a challenging experience for the VISC developers. The initial implementation plan included deployment of the application on both the Boulder and Denver campuses with the assumption that the IT support on both campuses will be similar. Implementation on the Boulder campus was a relatively smooth process. The DBA on the project was also the NT administrator and the network connections, database connections, installation of the runtime application, and connections to the images were all done by the DBA and developers on the development team. Accustomed to implementing administrative applications and providing all support around those applications, the group encountered a different IT support environment on the Denver campus when it came time to deploy the application there. At that time, the main support focus for the Denver campus IT department centered primarily on student lab support. Putting the application up on the Denver campus involved a tremendous amount of interaction between Boulder campus ITS staff, the Denver NT administrator, and the Visual Resources Center staff. The Denver customer support strategy had to be changed to include the VISC application support.

Future of VISC

This application was developed over a two year period and has been in production for two years. Given the rate of change within the software industry, it could be argued that this application is fast on its way to being outdated. The architecture of the system is two-tier client-server with a fairly heavy client-side application. The recommendation in the industry is to replace the heavy client-server application with a thin client, or at the very least with a three-tier architecture that allows for a thin client on the first tier, application processing on a middle tier and server based processing on the third tier. The Oracle database needs to be migrated from version 7.3 to the newest version, 8i, and the screens need to be updated from version 4.5 to 5.0. Likewise, the reports and queries also need to be upgraded to new versions. Upgrading the application to new versions of software not only enhances and ensures performance, but it provides additional functionality as well. For example, moving the database from Oracle 7.3 to 8i allows the images currently being stored in a separate file system on the server to be stored directly in the Oracle database. Images can be stored in one format and rendered in whatever format is desired by the Oracle 8i database.

Of course, migrating an application involves funding and the availability of further funding remains to be seen. Various Universities and colleges in the country have expressed an interest in purchasing VISC. In the summer of 1999, the department of Fine Arts on the Boulder campus began the process of securing funds to obtain a modified version of VISC for their slide library. The modified version will likely be provided in the newer versions of the software. Based on this interest, the Department of Technology Transfer at the University of Colorado has obtained rights to the software and distribution of VISC to other Universities remains to be determined. As part of the University’s "Total Learning Environment" initiative, consideration is being given to the possibility of developing a general application based on VISC that would be more generic in nature and would enable a variety of campus departments to use it for their slide collections. As more universities and colleges across the country start developing IT infrastructures and applications to enhance the student learning and faculty teaching environments, IT departments need to be receptive to the idea of collaboration to accomplish these objectives.