Holehan's UI Reviews

(:toc:)

Executive Summary

The KGet download manager for KDE is being rewritten for the upcoming KDE 4.0 and its developers Urs Wolfer and Manolo Valdes were looking for someone to do a usability review of their work so far. Two KDE usability evaluators, Florian Graessle and Tina Trillitzsch, agreed to do a Heuristic Evaluation of KGet. Before the review, the reviewers researched the current user base of KGet and that of competing download managers and determined their goals and needs. Together with KGet's developers, a target user base was defined and the reviewers created a number of frequent tasks to guide their evaluation. During the evaluation the reviewers went through the interface separately, with the help of a few common usability principles ("heuristics"), noting down any problems they perceived for potential users. Later they agreed on a severity rating for each issue and devised a suggestion to solve it. The results were put into a wiki and presented to the developers in KGet's irc channel. The reviewers discussed the issues and suggestions with the developers, who then almost immediately began solving some issues, so that these could quickly be marked as solved in the wiki. Overall 37 usability issues were found, most of which have already been fixed.

The findings showed a lot of promise: the browser integration is very well executed and KGet often offers very good feedback when adding downloads. There are some very innovative ideas for finding files to download on a web page. Most negative issues were minor, and only really significant in the main window, especially concerning the presentation and management of download groups and download details.

Introduction

This report documents a usability evaluation of the application KGet, done by two usability reviewers (Florian Graessle and Tina Trillitzsch) using the "Heuristic Evaluation" method. The reviewers both went through the interface, guided by realistic user tasks - essentially acting as a user - noting any problematic issues, agreeing on their severity and offering suggestions to solve them.

The report aims at accomplishing three things:

  • describe how the evaluation was executed
  • summarize the findings in the form of usability issues,
  • and offer recommendations to solve the issues

KGet is a download manager, an application designed to allow users to download files from the internet in a comfortable way. It can be integrated into the Konqueror web browser, replacing its more limited download feature. KGet's most prominent feature is a download list where downloads can be queued, paused, resumed and restarted. Downloads can be added to the list in various ways, including a "drop target", a free floating icon that users can place anywhere on their screen and where download links can be dragged and dropped from the browser or other sources. KGet is a KDE application - it is based on the K Desktop Environment and uses the GUI widgets and framework of that platform.

The developers of KGet approched the reviewers looking for usability help with their application. KGet is currently being ported from the old KDE3 framework to KDE4, and the interface and backend have largely been rewritten from scratch. The developers wanted to know how their new design could be improved.

Methods

Research into target users, goals and frequent tasks

To enable the reviewers to realistically play the role of a user during the evaluation, it is necessary to determine the audience that an application is designed for and to construct meaningful tasks this audience would frequently execute. Before the evaluation, the reviewers performed web research into the current users of KGet and their usage patterns and goals. Especially helpful were the KGet mailing list, its bugtracking system and several reviews of the app by users. For even more clarity about the goals and needs of people who use download managers, they looked at similar competing applications and read user reviews. After this initial research, the reviewers interviewed the KGet developers about their intended target users, comparing them to the previous research. Using the agreed upon target user base and its needs, the reviewers developed a number of frequent tasks these users would do with KGet.

The identified target users can be subdivided into two groups. Users from group A have a relatively slow internet connection and/or dialup internet. They often surf while downloading and don't want the downloads to eat up all their surfing bandwidth. For this reason they either want to stop and resume downloads once in a while, or they want to throttle the download speed so some bandwidth is left for surfing. Group A users often download a small number of large files (e.g. linux distributions in iso files). They like to be able to schedule downloads for later, when they are not at the computer or even asleep. Group B users on the other hand frequently download a large number of smaller files (e.g. linux packages) they don't want to handle individually. Internet speed is not as relevant for them as for group A. Both groups like a tight integration with their web browser so they can easily and quickly add downloads to their KGet download list.

The following tasks that were developed for the evaluation:

  • You want to download new KDE4 packages from http://www.kde.org/info/3.91.php. They are scattered over dozens of single packages all listed on one web page. Add all files to KGet.
  • Display status of the downloads in KGet, then pause all your downloads.
  • You are looking for new wallpapers and have found good ones at deviantart.com. Now you want them to be automatically placed in a certain folder while they are downloaded (only the images).

Two other tasks were ultimately not used because the functionality they were supposed to evaluate was not implemented in the new KGet version yet. These tasks were related to scheduling downloads for later and to throttling the download speed.

Evaluation method

KGet was evaluated by doing a "Heuristic Evaluation". The two reviewers went through the interface twice, each separately. The first round was devoted to getting an overview of the application. The second walkthrough was guided by the frequent user tasks, and by a number of common usability principles (ten heuristics by Jakob Nielsen) noting down any obstacles they noticed. Afterwards both experts combined all the issues they had found, discussed them and agreed on severity ratings for each issue. Then both reviewers cooperated again to devise suggestions for solving each issue, clarifying some of them with visual mockups. The results were placed into a wiki for easier coordination and then presented to the KGet developers during an irc chat.

Evaluation settings

The evaluation was executed on a computer running Kubuntu Linux version 7.04 ("Feisty Fawn"). The desktop environment used was KDE 3.5.7, which was delivered with the distribution. The KGet version used was checked out from the Subversion repository on July 23, 2007, as was the Konqueror web browser used together with KGet. Before starting each evaluation, any existing configuration files of KGet were removed to ensure a fresh installation.