THE UHLE COLLECTIONS @ BERKELEY

The Hearst Museum of Anthropology at the University of California, Berkeley houses an archaeological collection of ancient Peruvian artifacts known as the Uhle Collection. This collection consists of pots, potsherds, shells, beads, shells, textiles, human remains (mummies), among many other unidentified objects. Additionally, the Bancroft Library houses manuscripts, field notes and correspondence written by Max Uhle, the archaeologist that excavated the collection. Many of the items in this collection have been digitally photographed in order to allow researchers to view items in the collection without accessing the physical objects. A database has also been created to organize information about each artifact. This project provides a front-end to the database and digital collection. The goal for the project is to serve as a research tool by representing the relationship between artifacts and the metadata of each in a way that is useful and seamless to archaeologists. For instance, location is important to several archaeologists throughout the research process. With this in mind it would be useful to easily access location information at any point in one's search.

2. Solution Overview

To meet the needs of our target users, we have implemented a web-based interface to the collection of artifacts excavated by Uhle. The interface provides three access methods to the artifacts: keyword search, browse, and map-based browse. We have addressed the archaeologists' research needs by creating two views of our results-thumbnails to display shapes, and text-only to compare metadata. An artifact details page allows the researchers to view larger images and to access additional metadata. Furthermore, from the artifact detail page, users may move laterally through the collection. For example, a researcher looking at a particular artifact can expand the search to all artifacts from the same area.

Through these design elements, archaeologists will be able to gain access to the Uhle collections from all over the world, where previously access had been restricted to those in physical proximity to the artifacts. In addition, the interface serves as an easy way to compare artifact metadata across the collection.

[Top]

3. Personas and Scenarios

Personas

Initially, we had created three personas to represent three different types of users. We found, however, that one persona did not comfortably fit into the picture we were trying to create for the interface. As a consequence, we focused on designing for the remaining two personas, included below.

Susan Marks

Susan is a 55-year-old professor who has been an archaeologist for 30 years. Her specialty is Incan ceramics, and she is particularly interested in both the material composition and the iconography of the artifacts. Susan considers herself to be "not good with computers," but simultaneously prides herself in her ability to keep track of the latest archaeology technology. Software with a steep learning curve just irritates her. She is more inclined to farm out computer work to the several PhD candidates that she advises. Computers (and, perhaps, graduate students) are tools to use in her research-she has no interest in them for themselves except as they may further the work in her field and dislikes when they waste her time with extraneous tasks. Furthermore, she keeps paper copies of all her notes and important documents in filing cabinets because she does not trust computer media to be persistent. Thanks to her many years of experience, she is well-connected to other archaeologists and professors in her field, so she is confident that she can speak with some authority on many aspects of archaeology-from field methods and site administration to chemical analysis. She lives alone, but spends more time in the field or the lab than at home. Her office is filled with archaeological artifacts and knickknacks from her work abroad, but it is not cluttered. She likes everything to be just where she wants it, and gets upset if it takes her more than a minute to find an item that she has set aside. On the side, Susan enjoys working with pottery-not only is it soothing, but it gives her some insight to pottery methods used by the ancient peoples whose works she studies. In addition, she loves to travel and will jump at any chance to attend overseas conferences. Susan's Goals: To get her research done without having to focus on the tools or technology itself To make a difference and stand out from other researchers in her field To make contacts all over the world-expanding her information network through people

 

Persona #3 Rachel Richardson

Rachel Richardson is an undergraduate studying Anthropology with a focus on Archaeology at the University of California, Berkeley. After high school, she took a couple years off before deciding to return to school. She is now 23, in her junior year and hooked on school. She is starting to research graduate programs but is unsure whether her focus should be on ceramics, bone or soil. She has traveled to South America a few times and would love to travel more. In the past she was never much of a computer user but being in school has made her realize the value of email. She now uses email and instant messaging to keep in contact with family and friends, including some she met in South America. She is still a little fearful of the computer but fortunately she shares an apartment in the Mission with two roommates, one of which is very tech savvy. When she is not at school or a Mission taqueria she loves to read from her extensive library of books. She loves to collect things that tell a history, for example, she has saved every Muni fastpass since she moved here from Boston when she was 18. She enjoys cooking vegetarian meals for her roommates and friends. Rachel's Goals To get a clearer focus on her academic specialty To get into a good graduate program that will afford her travel opportunity To someday live in South America

Scenarios

These are the two primary scenarios that were developed from the above personas. Variations on these scenarios were used for our user tests.

Scenario #1: Rachel's Fieldwork Opportunity

Rachel knows that she needs to gain some experience in archaeological fieldwork if she is going to pursue a career in archaeology. While looking up possible sites, she discovers an opportunity to volunteer on a dig this summer near Chincha Valley in Peru. She remembers from her Andean Archaeology class that parts of the Uhle collection came from Chincha Valley, so she checks out www.uhlecollection.edu, a website her Professor recommended as a great information source with amazing images of the collection. She browses through the online collection looking at photographs of artifacts. She is not searching for anything in particular, but she is fascinated by the images and daydreams of excavating similar artifacts herself. Maybe this summer experience will inspire her to choose a specialty for graduate school.

 

Scenario #2: Susan's Pottery Analysis

Susan is in the midst of an ongoing research project into Incan pottery-specifically, she has been working on a chemical analysis of ceramics to see whether they were used for cooking, storage, or as decorative objects. In the course of her research, she has discovered that Berkeley houses many Incan artifacts, including ceramics, excavated by Max Uhle. She has used the Uhle Collections @ Berkeley website in the past to do some preliminary research, and is now at the stage when she needs to analyze the artifacts in person. At her office computer, she brings up the website for the Uhle collections. She searches for ceramics that appear to be burnt-possibly in a cooking fire. Susan browses through related artifacts and selects several that she thinks would be ideal for the first steps of her analysis. She prints a list of the selected artifacts with the appropriate museum number so that she can locate them in the physical collection, as there are no computers in the basement which houses the artifacts. Finally, she calls the collection curator to arrange a time when she can visit the collection.

[Top]

4. Final Interface Design Description

Functionality

Our final prototype allows a user to search through the Uhle Collections using three options: keyword search, map-based browse and category searching. Each additional search option adds a new parameter to the search results. Users are able to add and remove search parameters as they wish. We provide a thumbnail view, a text view and an artifact detail view.

Interaction Flow

View Larger Diagram

Not Implemented

Initially our system was to have a user log in option. This feature was not integral in the interaction flows of searching. This is a feature that will be implemented at some point in the life of the project but not during this stage of user interface development.
Our map has limited functionality due to a lack of specific information data in the collection. Some items provide region information but not all. We hope that the future of this collection includes refining this data so that the map can become more of a functional piece of this interface. Currently users may select a region to add a location parameter to the search.

Also, we have only implemented a subset of the checkboxes. Artifact Class, Material Type, and Location are all functional. However, we did not have data for Physical Attributes, Time Period, or Documents in our database. The facets are included on the interface to show the intended functionality, but they do not in fact work.

Finally, the "Total Matches" number is not active. This is merely a result of time constraints for this project-we chose to focus on making the checkbox and search items functional and were therefore forced to relegate the "Total Matches" number to a the bottom of the list.

Prototype Tools

Low-Fi Prototype

Tools: Our prototype was primarily paper based, with some features using transparencies, printouts and Post-Its.

Pros:

• Speed in creating - This prototype was really easy to build and fun.
  
  
• Spontaneous fixes - During user testing we were able to provide spontaneous solutions to unanticipated actions.

Cons:

• Slow "computer" - While testing our human computer was slow to respond but out participants were patient.



• Hard to test all features - Using paper and markers it was hard to tell if users would really notice some of the changes once the prototype was web based.



• Paper cuts - Ouch.
  

Interactive Prototypes (First, Second and Final)

Tools: These prototypes were built in HTML and JavaScript and use CSS. We used the following software: Macromedia Dreamweaver, Adobe GoLive, Adobe Photoshop, TextPad, TextWrangler, and Taco HTML Edit (text editors for the Mac). A pre-made script from http://www.dynamicdrive.com was used for the menus, albeit modified extensively for our use. The database is written in MySQL and is connected to front end search features using PHP.

Pros:

• Reusable code - Using HTML and JavaScript reduced future coding by building on the base we have written.

• Realistic representation - Using these tools through the iterations of the prototype it was easy to simulate what features might look like when fully implemented.

Speed in implementing - Our team had experience using these tools so coding time was reduced.

Cons:

• Many error pages: Through the design and testing process we did not fully implement duplicate features to save time. For instance we did not make all of the checkboxes active in the first and second prototypes. However during testing our users managed to find a lot of those unimplemented features.
  
  [Top]

5. Design Evolution

The following sections describe the design changes we made throughout the course of the semester, from rough initial diagrams on paper to a working interface on the computer.

Initial Sketches

After we carried out preliminary interviews with primary users and a comparative analysis with other search interfaces, our initial designs were focused on three different methods for accessing the information in the Uhle Collections database: Browse, Search, and Map-based Browse. For this phase of the design, we developed each method on its own to better understand the strengths and weaknesses, although we knew that elements of all three designs would be incorporated into the next iteration of the design.

The browsing-centered design allows users who have some prior knowledge or experience with archaeology to explore the collection along one of four top-level categories: Location, Time Period, Material, or Artifact Type. Also, making the categories visible allows relative novices to get a broad picture of the collection's contents as a whole.

A "Google-like" approach was applied to our search access point. It allows the user to enter a keyword in the main page and immediately bring up a list of results relevant to the search term.

Lastly, in response to preliminary user interviews, we proposed a map-based method for accessing the Uhle Collection data. Through this method, the user has an excellent visual representation of where artifacts were discovered in relation to each other and their surroundings.

Low-fi Prototype

The purpose of the first prototyping experiment was to determine whether our initial design ideas were natural and easy for our prospective users to use without prior training. We had combined the all three methods from our initial design. We based part of our model on Flamenco's tightly integrated search-and-browse function, but geared conceptually toward artifacts rather than images. We also wanted to test whether map-based browsing was useful.

Our low-fi prototype was primarily paper based, with some features using transparencies, printouts and Post-its. The basic layout of our prototype involved three main sections: title, navigation, and display. The title displayed the project name across the top portion of the screen and does not perform any function. The navigation located in the left-hand side of the screen was composed of three ways to search the collection: keyword search, interactive map, and browsing top level categories. On the right-hand side of the screen, the main content of the page was displayed, such as information about some of the artifacts from the collection in the opening screen, or search results with different presentation options like thumbnail images or text only views which were added on later testing.

After our first user test, we found out that every user liked the checkboxes that displayed all the available metadata they could use for browsing when beginning their search. We also learned that users had a tendency to navigate to "Related Document" through the artifact detail screen rather than through in the top-level "Documents" category. Most critically, we discovered that there was some confusion among the archaeologists about terminology. There was overlap conceptually between some terms and many questions about proper classification were raised throughout the tests.

First interactive Prototype

With our first interactive prototype, we tried to address the many issues that were brought to our attention from low-fi prototype tests. The framework of the first interactive prototype was built in mainly HTML and JavaScript. Because the backend database for the collections was undergoing extensive reworking, we implemented temporary mock-up search results.

Overall, our first interactive prototype remained quite similar to the low-fi prototype in terms of basic layout and navigation. However, based on user testing of our scenarios using the low-fi prototype, we made changes to the following items:

Title links: The information about the site and project as a whole moved under the title and remained constant when users navigated through the collection.

Text only view: Similar to an Excel spreadsheet of the artifacts, this allowed a user to easily compare metadata between artifacts.

Artifact indicators: Inspired by Flamenco, these numbers represented the number of items remaining in the collection when that facet was checked. In our low-fi prototype, users were confused by the sheer "number of numbers," so we chose to simplify this as a single number at the top of the category menus and added a "Go" button so that users had more control of the amount of displayed results.

Heuristic Evaluation and Second Interactive Prototype

In our second Interactive Prototype, we refined the system interface based on the Heuristic Evaluation we received from the PhotoCat team. Much of the feedback was valuable and some of the changes are briefly outlined below.

Function Heuristic Evaluation by PhotoCat Uhle Solution

Function	Heuristic Evaluation by PhotoCat	Uhle Solution
[H7 Flexibility] (Severity 3)
Search	The gray background combines three forms with three different "Go" buttons.	Separate color sections to distinguish the separate areas.
[H1 Visibility] (Severity 3)
Search	Confusion of 'Show All" button.	Added the ability to expand all the menus and close all the menus.
[H10 Help and Documentation] (Severity 4)
Help	No help visible from for the system.	Added a help page to assist users.
[H5 Error Prevention and H6 Recognition rather than recall] (Severity 3)
Browsing information	Unclearness of "Remove link" option.	Created a colored box that surrounds the entire parameter set so that remove is more easily associated with the grouping.
[H7 Flexibility and efficiency of use] (Severity 3)
Browsing information	The position of the "views/thumbnails/text only" menu and visibility.	Changed the layout of the thumbnails/ text only views. Also implemented radio buttons that indicate which the user is looking at.

 

Final Usability Test

We conducted the final usability test with our second interactive prototype. Because our users tended to be very exploratory and because our prototype was limited enough that tasks tended to run together, we focused on qualitative rather than quantitative data for this round of testing. Even though our users rarely made "errors", we had to address several issues in our final prototype. The most important changes we made to the prototype as a result of the final usability test include the following:

• Improving graphic design to more effective and efficient use of interface
  
  
• Improving map-based browsing method
  
  
• Connecting to working database

Usefulness of the Evaluation Techniques

Our final prototype has come a long way from our initial design, and the various evaluation techniques helped us turn our dream project into reality. Each evaluation method brought with it different issues or information that we needed to improve the next iteration of the prototype. Preliminary interviews with primary users gave us an array of possible ideas about basic layout and functionality. The low-fi prototype yielded feature and terminology refinements. Both the first and second interactive prototypes brought to our attention the overall usability of the design.

Among all these useful evaluation techniques, the low-fi technique was the most valuable because it was the first time we could test with actual users. Because of the unrefined and clunky look and feel of the paper prototype, users were likely to be more forthcoming about the problems they saw, and to suggest alternatives. The issues that users brought to our attention remained as major themes for us to work with to improve our system throughout the project, such as the constant reworking of terminology for proper classification. Such rich results were not as evident in later testing cycles, although we believe that further work will yield greater refinement.

[Top]