This week was mostly spent on further research into UI.  I thought I’d take this opportunity to present a screenshot of where the VISTAS UI is at today.  It is currently a “bare metal” design; functionality has been kept at a minimum while the rendering engine has been worked on.

The Vistas UI, as it stands today.

Upon recommendation from Judy, I took a look at another visualization program called VisIt.  VisIt has been under development for many years by the Department of Energy, and in that time has become very feature-rich.  As you can see from the following screenshot, it is anything BUT “bare metal”:

A screenshot of VisIt, another visualization program.

As the scope of VisIt is presently much larger than VISTAS, I highly doubt VISTAS will ever need anything close to that many UI elements.  Nonetheless, I thought VisIt had some interesting features.  Namely:

• Separate windows for loading/project management & visualization
• Axis locking and zoom controls
• Ability to change colors used dynamically via color table
• Camera views – ability to lock, save & load
• Save/Print Window for easy export of imagery

I also read an interesting paper: Current practice in measuring usability: Challenges to usability studies and research, Kasper Hornbaek, International Journal of Human-Computer Studies, Int. J. Human-Computer Studies 64 (2006) 79–102.  In it, Hornbaek acknowledges and evaluates the efficiency of various metrics for usability.  Particularly, he identifies the following measures of usability which can prove useful to my work on VISTAS:

• Binary Task Completion — did the user finish their task?
• Accuracy — how many errors did the user make while working towards their task?
• Recall — how much information can the user remember after using the program?
• Completeness — how complete of a solution did the user achieve?
• Quality of Outcome — the user’s subjective satisfaction with the end result of program use

The differences between Accuracy, Completeness, and Quality of Outcome are subtle but very important.

In other project news, we are having a VISTAS teem meeting in Corvallis next week.  I am excited to visit with everyone, and am looking forward to the opportunity to talk to Margaret Burnett, an OSU professor who is an expert on UI and does some really cool research into machine learning techniques applied to UI.

I spent this week cloning the latest VISTAS build off of the bitbucket repo and making sure that my Visual Studio can successfully build it.  This was a bit tricky, as large open source software projects inevitably gather a lot of dependencies, but I believe that I have my workflow in place.

I also set up this blog, submitted my project proposal (see previous post) and did some online resource hunting.  I’ll list these resources here for me to come back to…

VISTAS Related:

VISTAS Project Blog
VISTAS Source Repository

wxWidgets Related:

wxWidgets Alphabetical Class Reference
wxWidgets Guides & Tutorials
wxFormBuilder – Open Source wx Designer

Research Papers of Interest:

User Interface Design with Visualization Techniques – A pretty basic paper, but introduces an interesting scoring rubric for UI analysis
Current practice in measuring usability: Challenges to usability studies and research – More interesting thoughts on UI quality heuristics

Summary & Learning Objectives

For my senior project, I will be working collaboratively on a small development team working on VISTAS, a 3-D scientific data visualization tool.  This project is taking place under the mentorship of Judy Cushing and is funded by the National Science Foundation.  VISTAS has already been under development for several years, so a big challenge of my project will be learning my way around an existing large C++ codebase.  The rendering engine in the application is mostly complete; however, the tool is currently slightly buggy and can be somewhat unintuitive to use.

My role will be to focus on learning user interface design & programming with the goal of enhancing the usability of the program.  I also hope to learn about modern source control methods as well as obtain practice at working on collaborative software projects.  Throughout the course of my project, I will also be fulfilling additional developer responsibilities within the VISTAS team, including attending weekly online meetings and contributing to code reviews, bug fixes, and collaborative readings of select research papers.

Resources

The existing VISTAS UI uses a cross-platform library called “wxWidgets” for all of its graphical UI elements.  There are official wxWidgets tutorials as well as a reference page online– these will be my primary resources for this project.

VISTAS-related resources can all be found at the project’s webpage.

Implementation Plan

My plan initially is to target making addition and removal of visualization scenes more intuitive.  I also hope to add context-driven mouseover tooltips for user interface elements and a basic in-application help menu complete with a quick-start guide.

When it comes to exactly what I mean by “making addition and removal of scenes and visualizations more intuitive”, I had in mind:

• Add ability to delete items from Data & Visualizations tabs.  Currently, removing a scene requires rebooting VISTAS.
• Add “Add” buttons to the Data & Visualizations tabs.  Right-clicking is not intuitive to a new user.  This could be simple as a small “+” or “Add” button next to Project Visualizations/Project Data.

Beyond this, I’ve compiled a list of some user quality-of-life features that I think would be good additions.  The big two in my mind are the ability to Save/Open/Close projects and the ability to add annotations/pins to visualizations, but I’d like to work up to these.

Evaluation of Work

All of my code commits to the project can be viewed at my bitbucket profile (https://bitbucket.org/jmangue/profile/activity).  Furthermore, I will be keeping weekly online journals detailing my work on VISTAS.