Assignment 2, Due Thursday, June 21st

  

Part 1: Installing and Getting to Know CogTool

If you haven’t done so already, please download and familiarize yourself with CogTool in the way described in the slides from Tuesday, June 5th. You may find it helpful to work quickly through the tutorial example that we went through on June 5th to get used to some of the quirks of the CogTool interface before starting to use it creatively. At various points, you will find it helpful to refer to the User Guide; this Guide is very lengthy, but it is well indexed, so you can quickly find the information that you need.

As was announced at the beginning of the class of Tuesday, June 5th, the explanations in the following instructions assume that you have attended the lectures of Tuesday, June 5th and Wednesday, June 6th. These lectures included highly interactive demonstrations and discussions of CogTool and the prototype discussed below which it would be impossible to capture in slides or written notes. If you skipped or missed one or both of these lectures, please fill in the gaps in your understanding by consulting with another member of your group or another student in the course who is willing to help you. 

Overview of the Logic of This Assignment

The goal of this assignment is to experience what it is like to explore a new general interface idea using a medium-fidelity prototype. We introduce a new style of browsing called parallel faceted browsing. This browsing style is based on a few core ideas, but it can be realized in many different ways with a user interface. Also, there are various conceivable uses to which the interaction style can be put, and no-one yet knows what other uses might still be thought of or how useful any of these really is. You will start with one possible realization of this idea in an interface, which you will have an opportunity to improve on either now or later. You will look for a domain and types of task for which paralleled faceted browsing seems to have added value for users. You will prepare some material that can be used in the next assignment for testing your hypotheses concerning these questions. 

Conceptual Background: (Traditional) Faceted Browsing

Faceted browsing (also called faceted search) appears in various forms in many contemporary websites. (The topic is not covered in our textbook, but you can find quick introductions in places like Wikipedia.)

The basic idea is that there is a large set of objects that a user may be interested in. These may be, for example, news stories, products in an electronics store, flights that can be booked, or works of art that can be admired. Each object can be described in terms of its values on several dimensions, or facets. The user essentially starts with the entire set of objects and narrows it down by specifying one or more values for a facet. This filtering process is repeated until the set of objects is a smaller set that the user is interested in – for example, the set of works of art that were created by a particular artist in a particular genre within a particular period of time.

A limitation of traditional faceted browsing is that the user can zoom in on only one subset of the total item set at a time. If the user wants to consider a different set instead, he or she can do so only by specifying different values of facets, thereby losing sight of the subset just focused on. For example, if you have found the set of all flights from Saarbrücken to Heraklion on the morning of June 11th and then decide that you might get better results with flights from Cologne on the afternoon of June 12th, you have to specify different facet values. It is therefore not possible to compare the two sets of results (except by using ad hoc methods such as opening a second browser window or making a screenshot of the first set of results). 

Parallel Faceted Browsing: Initial Mockup

The basic idea of parallel faceted browsing is that the user should be able to specify multiple combinations of facet values and have simultaneous access to the set of objects that corresponds to each combination. And since the user may want to specify a considerable number of combinations of facet values, it will be important to organize them in a way that makes visible the relationships among the combinations.

The method chosen in the initial CogTool prototype made available for this project is to enable the user to grow combinations facet values like a tree, from top to bottom. This procedure is a little bit like that of moving down a hierarchy of events from the top node in the GLOCAL prototype.

To see an illustration of this idea, please download and open the CogTool project called PFB and walk through the demonstration of the task called “Check Three Days” by stepping down the list of actions in the right-hand pane with the “Down Arrow” key.

As you will see, the screen is divided into two parts:

On the left, you see four facets, each with three possible values. On the larger right-hand side, a tree is built up incrementally. Each noninteractive button in the tree (without a border) stands for a facet or a flight. (A flight is displayed simply in terms of its price in euros; in a more complete prototype, the user could click on the price to get a detailed description of the flight.)

Each button in the tree (with a border) stands for a combination of values of facets. At any given moment, one button is in focus, as is indicated by the “>” sign in front of it.

The user builds up the tree starting from the topmost node, which is initially in focus. When the user clicks in the left hand side on the button for a facet value, the tree is expanded downward to add to that facet value below the selected node, which then becomes the focused node. The user can put a different node in focus by clicking on it.

In this way, the user can specify several combinations of facet values that represent interesting subsets of the total set of items. To see a list of the actual objects that belong to each subset, the user can double-click on the lowest button that specifies that combination of facet values. 

Part 2: Extending the Initial Prototype

To check your understanding of this prototype (and of CogTool itself), first extend the initial prototype a bit. Add enough new frames and transitions to make it possible for the user to check out flights from Cologne on the afternoon of June 12th or 13th in the same tree.

Hints: When creating a new frame, “duplicate” the most similar existing frame and then edit the new one. You can add items to the tree on the right by copying and pasting the relevant existing items from the left-hand side of the screen. To add a noninteractive button for a particular flight, you can copy one of the existing similar buttons and just change the price.

Create and “demonstrate” a new task called “Consider Cologne as Well” that starts in the same way as “Check Three Days” but then proceeds to add the check of flights from Cologne mentioned above. 

Part 3: (Optional:) Choosing a Different Example Domain

If you can think of another domain that interests the members of your group more than the domain of flight search, you can make a new prototype like the existing one that refers to this domain. You can then work with that domain from now on. An example would be the domain of recipes. It will be interesting if some groups work with other domains, because each domain may suggest interesting design ideas.

Hint: You can create the design for the new domain in the same CogTool “project”, using the “Create” / “New Design” command. It will appear as a new column in the main spreadsheet. To make the first frame of this design, copy the first frame of the existing design “Initial Design” and replace the labels of the various buttons with labels that make sense for your domain, perhaps changing the number of buttons on the left-hand side of the screen. You can then quickly create new frames in the way described above.

Create and demonstrate at least one task in your new domain. 

Part 4: (Optional:) Design and Implement Changes in the Basic Procedure for Operating the Prototype

The realization of parallel faceted browsing in this prototype is just one of many possible realizations. We are not interested at this point in variations that concern the detailed screen design (e.g., the colors of the buttons or the exact layout on the screen). But if you have an idea about a fundamental improvement, you may want to realize it now, so that you can continue to work with an interface that you like better. An example of such an improvement would be to allow two or more nodes to be in focus at any given moment; in that case, when the user clicks on a facet value on the left, that facet value is added underneath all of the nodes that are in focus. (In this way, a user could conveniently specify for each of several days that he or she is interested only in leaving in the afternoon.) It may require some thought to figure out how your improvement can be realized in the interface. If you decide to stick with the existing interface for now, you will have opportunities later in the course to explore improvements. If you do choose to change the basic procedure in some way, please add a paragraph or two explaining your change in the document described below. 

Part 5: Preparing Tasks and Stories That Suggest the Added Value of Parallel Faceted Browsing

In the next assignment, you’ll be showing your prototype to potential users with the goal of finding out whether parallel faceted browsing offers added value to them. So your goal now is to think of tasks for which you believe that parallel faceted browsing may be especially useful. (See also the “Preview of the Next Assignment” below.)

Ideas About Possible Benefits of Parallel Faceted Browsing

1. When what the user wants to do is choose a good object from a large set of objects, this paradigm allows the user directly to compare different subsets of objects instead of examining one subset at a time.

2. In the same situation but a bit less obviously, the user can start examining different subsets in parallel without knowing in advance which ones will be promising. At any given moment, the user can focus on the subsets that seem most promising and expand these further. In our flight example, the user might start exploring all three departure airports; if after narrowing each of them down a bit user discovers that the flights from Saarbrücken tend to be much more expensive than those from the other airports, the user can stop examining Saarbrücken further and concentrate on the other two.

3. It is often interesting to be able to compare subsets of objects even if your goal is not to choose a single object. For example, if the objects are famous paintings, the user might find it interesting to compare the paintings that two different painters created early in their respective careers.

4. The user may want to make two or three decisions in parallel because they are interrelated. For example, the user may want to plan a brief vacation in June and a longer vacation in August. Examining possible flights in these two months in parallel has the advantage that the relationships between these two trips can be taken into account. For example, the user may want to fly to the same place in both months, though he she doesn’t yet know what place that is.

What to Prepare

Mock up two tasks that seem to illustrate the added value of parallel faceted browsing in your domain and that seem suitable for presentation to potential users in a later phase.

For each of these two tasks:

1. Write a textual story that explains how the user proceeds to execute the task with the prototype. Example: “Petra wants to take a brief vacation next week to some warm place in southern Europe. The possible airports she could depart from are Saarbrücken, Zweibrücken, and Cologne. Petra starts by looking at the flights from Saarbrücken and Zweibrücken ....”

2. Write at least one paragraph explaining why you think parallel faceted browsing is helpful for this task. 

What to Submit

1. Create a folder with a name like “A2 Schmidt”, where the name is the name of the group member who submits the report. Share this folder (only) with Tanja Schneeberger (tanjaschneeberger89@googlemail.com).

2. Into this folder, please upload the .cgt file of the CogTool prototype that you have created as an extension or modification of the one provided above. (It will be stored in Google Docs as a “document” that consists only of a link to the prototype.)

3. Into the same folder, upload the text document that contains your stories and explanations. 

Appendix: Preview of the Next Assignment

To give you a better idea of what you are trying to achieve with the previous steps, here is a preview of what you’ll be doing a bit later, starting in the week of Monday, June 25th:

1. You will present your mockups and stories to people who might be interested in using the system in question.

2. You will observe them as they walk through the mockups and read the stories, and you will ask them a combination of closed and open questions. Your goal will be to find out (a) whether parallel faceted browsing actually has the expected benefits in this situation; (b) whether it has unexpected benefits or drawbacks; (c) whether there are additional requirements that the interface should meet that you haven’t thought of already (aside from the obvious requirement that it must look nicer and be easier to understand).

3. You will analyze the collected data and discuss it in terms of its implications concerning possible realizations of parallel faceted browsing interfaces and possible benefits and drawbacks of the parallel faceted browsing paradigm. 

What to Do If You Have a Question

If, while working on the assignment, you have a question that cannot be answered with the help of the instructions or the other available material cited above, please send it by email to Anthony Jameson (jameson@dfki.de). Unless the answer is very simple, it may not be sent until Sunday, June 17th or shortly afterward. Note that I will not be able to repeat by email parts of the lectures in which this assignment was introduced.