UAN and Cognitive Walkthrough

Comments on Glasgow Solutions

Phil Gray

2 March 1998

 

General Comments on UAN

  1. The task should be described in a set of tables, one for the top-level task and others for the subtasks. You can use a "task tree" to represent the goal hierarchy graphically, but then use UAN tables to specify the nodes of the tree in detail.
  2. UAN has four basic columns, viz., user actions, feedback, system state and application operations. None of your UAN descriptions made use of this final column. It will have one entry, recording the actual change in the column width, which is an operation on the underlying document and not just a user interface operation.
  3. In specifying column selection, several groups included an intermediate action, (~[x,y])*, between selection of the diagonal corners. In UAN you do not have to indicate as separate user actions all of the intermediate cursor movements. All the user must do is move to the uper left-hand corner, hold down the mouse, and then move to the bottom left-hand corner, i.e.,
      ~[top left corner of col 1] Mv

      ~[bottom right corner of col 1] M^

  4. The column labelled "System State" (or "UI State") is intended to include state, not otherwise captured in the user actions themselves, which can affect the consequences of user actions. Sometimes these states are referred to as "modes". In this example, the only significant states are the selected column, which affects the contents of the subsequent dialogue box, and (perhaps) the activation of the dialogue box itself. Selecting a menu by mousing down on its label creates a state which is indicated by the description of the user action as selecting the menu item by name; including this in System State isn't wrong, but is unnecessary.
  5. The UAN analysis can be used as the basis for your cognitive walkthrough. You don't need a separate analysis called a 'Goal Hierarchy'.

General Comments on Cognitive Walkthrough

  1. The purpose of a cognitive walkthrough is to identify potential usability problems. The input to the walkthrough is a scenario, which will usually be incomplete, especially with respect to information about what the user already knows. Often, one won't know for sure, but can hypothesize a possible problem. For example, we know that Kim has used Macs and Word before, but we don't know if she has ever created or formatted a table. Therefore, it is possible that she doesn't know how to select a column nor the relevant menu item for column resizing. Some of you caught this potential problem, but several groups did not.
  2. One of the most likely problems to be encountered is the fact that Kim must be very lucky to pick the right value after only two tries. There is nothing in the interface to indicate the correct value and no way of determining it other than by trial and error.
  3. Some of you weren't sure how to handle alternative methods of performing the task other than that described in the scenario. Cognitive walkthrough depends on the scenario description. There may be other ways of carrying out the task, but these are not handled by the CW technique given here. This is a possible weakness of this evaluation method.
  4. In writing up your cognitive walkthrough, you should include (i) the questions you asked and the answers you gave and (ii) a summary report of problems found. The first section provides the evidence for the problems in your summary and also indicates what you believe will not be a proboem (relevant, too). The second section is important because finding and dealing with these potential problems is the aim of the walkthrough and a summary makes the walkthrough much more usable.
  5. Giving a likely level of seriousness for each problem can help in using the cogntive walkthrough results, e.g., in prioritising the problems for further investigation.

     

 

Comments on Group Solutions

Group A

Group B

Group C

Group D

Group E

Group F

Group G

Group H

Group J

Group K

Group L

Group M

Group N

Group P

Group Q

Group R

Group S

Group T