ZBento Design Meeting, 2002
ZBento Design Meeting, 2002 |
My experience in web development is quite extensive, reaching back to the early days of Mosaic, when I created a web site for my college newspaper comic strip, "Ka-Blouie!". As web technologies progressed, I acquired skills in CGI script programming and eventually co-founded a web design company, Yeti Arts. Then came XML and content management systems, such as Zope. In the two years immmediately preceding my term in the Information School MLIS program, I worked for the University of Washington Program for Educational Transformation Through Technology (PETTT) as a web site manager, web applications developer, and educational technology researcher.
When I began my graduate studies at the iSchool, I continued to refine my skills in IT design and development, adding competencies such as user-centered design, knowledge management, and object-oriented programming. I found new ways to incorporate skills learned in classes into my projects and allowed the theory learned in my studies to inform practice. Below, I explore and reflect on my most significant experiences in which I participated in the design and development of projects and products involving information technology.
 iSchool Research Event Poster: Faculty Accomplishments |
In Summer 2002, iSchool instructor and content management guru Bob Boiko asked me to be his graduate assistant. He told me that he was working on a project involving the design of a university-wide information system for recording and tracking the achievements of faculty at the University of Washington. I had taken a class Bob taught earlier that year, enjoyed working with him, and had liked reading his book The Content Management Bible, so I accepted, hoping that this experience would be positive as well.
My role was to be the "Information Lead", in charge of user analysis, information architecture, and user-centered design. The programming was to be carried out by Suzi Soroczak, a PhD student at the iSchool. The project was dubbed "Faculty Accomplishments".
I had little formal training in user-centered design (UCD), so my first assignment was to scan the literature on systems analysis, needs analysis, qualitative research methods, usability testing, and so on. Over the months, I amassed a library of UCD manuals.
The Faculty Accomplishments project was mandated by the UW administration with support from a couple of faculty from the Information School and School of Nursing. Our goal was not to build the information system, but to develop a conceptual design that would be submitted for consideration at the end of the year.
We developed a strategic plan during Autumn 2002, exploring the best approach to take, and by the end of the quarter, we were ready to implement that plan. Bob worked on database schemas and algorithms, and Suzi reviewed the technical requirements of the university systems. Meanwhile, I began interviewing faculty and staff members in the Information School and the School of Nursing.
Over the next 4 months, I interviewed 13 people, including faculty members, deans, chairs, support staff, and student assistants. Using the information I collected, I then produced systems and user analysis documentation, including:
 Faculty Accomplishments Information Architecture Diagrams |
My goal in doing all this analysis was to obtain a clear conception of the needs and motivations that govern the workflow of recording and reporting on faculty activities. By using that conception to design the system, the Faculty Accomplishments system will fit, hand-in-glove, into the users' existing tasks and systems.
Some of the most significant findings from this process could probably not be elucidated by common technology-based system design methods. We had originally underestimated the role that staff and students currently play in recording and reporting faculty accomplishments. In fact, it seems that staff and student assistants will be the most frequent users of the Faculty Accomplishments system. I also found a great need for an incentive structure (such as promotion or penalties) to encourage -- or coerce -- faculty to regularly add and update their information in the system.
 Faculty Accomplishments Wireframes |
I see this project as one of those rare opportunities to design a system from scratch -- and do it right. I have worked on several technical system development projects in the past, but was often rushed to build a system before it was ready. By taking a metered, thoughtful approach to user and systems analysis, we have been able to enter the design phase with very little ambiguity or guesswork.
Through the Faculty Accomplishments project, I have gained many valuable skills -- the most important of which include:
Unfortunately, this project never came to fruition, due to a lack of funding or support from the university administration, and was tabled indefinitely. However, it remains one of my capstone learning experiences as an information architect.
 zBento data model & system architecture |
I have worked continuously for the University of Washington Program for Educational Transformation Through Technology (PETTT) in some form since October 1999. During my time at the Information School, I've worked as a 1/4-time graduate assistant for PETTT. Over the last nearly 4 years, my role in PETTT has changed from web site manager to educational technology researcher to web applications developer to system designer. The theme of all these roles has been consistent, however: knowledge management (KM). KM is a buzzword that usually means using a set of techniques and tools to capture, store, maintain, share, and utilize knowledge (information and the stuff in people's heads) within an organization that values learning.
PETTT has assisted several entities on the University of Washington campus that develop tools and strategies for educating learners, in classrooms, in distance programs, and in the world at large. The tools, which eventually took the form of a gradually evolving content management system, were designed and developed primarily by myself. When I started graduate school, however, I passed most of the development responsibility to Eric Maddox, a full-time developer for PETTT.
During Spring quarter 2002, in which I took Classification Theory (LIS 535) and wrote a paper on Epistemology and Classification, I attended a meeting in the Architecture school to discuss how PETTT could help apply its content management tools in a study-abroad build-design class. We were discussing how all the information they were collecting would be organized, and I had a brainstorm: What if we built a content management system that used faceted classification? This led, eventually, to a strategy for creating prototypes, such as the Legal Gateway Project, and our development of a new content management tool called zBento.
 zBento facet search interface |
Over the next month, I mulled over this idea and finally presented it to the PETTT team. I explained how a system that was able to more faithfully represent knowledge in its organization (information architecture) and, hence, its navigational scheme, might be a better knowledge management system as well. They were excited about the prospects, so Eric and I began designing the system.
We started first with the data model. A faceted classification system is a poly-hierarchical, multidimensional subject listing, so our database had to represent those aspects. The content management system also had to provide for the classification of documents within that faceted system. During several day-long work sessions, we hashed out what the entities were, how they were related, and, eventually, what the logical schema of the database was. From those, we created UML class diagrams and built prototypes of the database and scripts that would generate navigational structures. Although Eric did the majority of the programming, we worked side by side on many occasions; I provided debugging and algorithm editing support. We also chose a name: zBento. The "z" is for Zope, the web applications development platform we used to build the content management system, and "bento" is the Japanese lunch box with compartments for organizing a meal.
Development of zBento went slowly, since PETTT had a number of other projects that both Eric and I were involved in. However, we managed to incorporate our prototypes into many of those projects, including the Legal Gateway project. Lessons learned from those prototypes then informed the design of zBento. While we were doing all this, another PETTT team member, William Washington, designed and conducted usability tests on a user interface for zBento. Over the next 8 months, we'll write all the logic code to marry the data services of zBento to the user interface. Soon, what began as a thought experiment in a class will become a reality.
 zBento management interface |
This work was very closely coupled with my intellectual argument that faceted classification can be used to generate information architecture. I wish that I had more time to devote to zBento and guide its progress to match my original conception of it. Unfortunately, technological and logistic limitations caused the project to change as time passed. However, the faceted classification system -- the core idea behind zBento -- remains integral to the project.
More than anything, I feel I have gained an understanding of the life cycle of a long-term software development project. The seed of an idea became a team effort with very real consequences to the users of the systems we built.
The skills I learned in Information Systems, Architectures and Retrieval (LIS 540), Database Design (LIS 542), Information Systems Design (LIS 543), XML Schema Design (LIS 598) constantly informed the technical decisions made in this project. Furthermore, the intellectual work and theory gained in Information Behavior (LIS 510), Organization of Information and Resources (LIS 530), and Classification Theory (LIS 535) gave me the inspiration and grounding for making zBento a reality.
I worked with the zBento team until June 2004. Several of the prototypes that informed the development of this project, including the Legal Gateway and the Architecture project, went live, and we observed how these prototypes were being used. Based on the results of these studies, we adjusted our design accordingly. Additionally, I and the zBento team presented our work at the September 30, 2003, WebEd Meeting and in an administrative review of PETTT.
Even though the administrative review went very well, the PETTT group was eventually dissolved due to political factors, and our work on zBento, save for the Legal Gateway and the Honors Program in Rome, never surpassed the prototype phase. Members of the zBento team went on to work for other UW units, and, hopefully, the knowledge gained from this project will be put to good use.
I get sick of building web sites sometimes. HTML is so boring and obtuse, and server-side includes alleviate only a little of the tedium. Yeti Arts, the web design company I co-founded with a few friends, has several clients whose sites I update on a regular basis. And when those sites are in HTML, it makes the whole maintenance process very unrewarding.
One weekend I got inspired to build myself a content management tool. So I sat down and pounded out a simple blogging tool that I named, imaginatively, YetiBlogSite.
Using everything I'd learned about web site and user interface design, content management, and programming up to that point, I composed the code in Zope (which uses DTML and Python) for a simple tool that would:
As evidenced by this YetiBlog-based portfolio, YetiBlogSite does all this... and not much more. And yet it is most likely one of the most useful tools I have ever built.
YetiBlogSite works so well probably because it is so simple. I created a blank, default YetiBlogSite for my brother, who knows little to no HTML and let him go at it. Within a month, he had blogs for his photos, his resume, and his journal (which is read frequently by his friends and family). It seems unlikely he could have done the same with zBento.
I think I learned that sometimes -- maybe most of the time -- less is more. I'm not sure how all this fits into this portfolio, but I felt that it was worth a mention because, as I mentioned above, this very site was built with YetiBlogSite.
I am constantly refining YetiBlogSite -- making it simpler, more reliable, and more intuitive. As with most things built in Zope, it's open source, so the export file is available for download and modification below.