Usability/Workshop/CurrentState

Z Kadu

Current state of the Kadu Usability Project as of 04.2009:

A document meant for Season of Usability mentors and students

--Akai 16:42, 26 kwi 2009 (UTC)

As you can read on the project's main page Kadu is an Open Source Gadu-Gadu protocol instant messenger for Linux, BSD and OS X. It is the most popular Gadu-Gadu client on Linux (probably on BSD too) and the most notable Open Source application created in Poland. Chances are, you can find Kadu in the repository of your favorite Linux distrubution.

Gadu-Gadu is the most popular IM service in Poland, with over 7.8 million registered accounts and approximately 6.5 million users online daily. Ever since its creation in 2001 Kadu has been a Gadu-Gadu - only internet messenger, meant primarily for Polish users of this protocol. Right now we're working on a new version of Kadu with support for more protocols, primarily Jabber but also others in the longer run. We hope to attract international userbase with Kadu as a Jabber client as well as make life easier for Polish users, many of whom use Jabber on a daily basis (according to our survey, that's the case for 30% of Kadu users). Along with adding support for multiple protocols, we're planning to support simultaneous connection to more than one communication network, using multiple accounts of the same protocol at the same time and metacontacts. All these enhancements require big changes in the interface of Kadu (which was not flawless usability-wise to begin with and may require additional changes even after the big redesign we're working on right now is over, especially in the configuration manager).

The Kadu Usability Project has been started in the summer of 2008 to provide developers with the design of the multiprotocol interface for Kadu, as well as to address other usability issues in Kadu in the future. The people behind it are me ( Joanna Pierożek aka Akai ) and Kadu's PR manager Piotr Pełzowski with minor contributions from other people and organizations mentioned here and additional help from various members of Kadu's user community and, of course, the developers.

The project infrastructure includes the wiki page: http://www.kadu.net/w/Usability where we're documenting the progress of our work and store other documents related to the project, a Kadu Design Feedback mailing list used for announcements, comments and communication with the developers and online survey software running on the web server of the KateOS Linux project that can be used to conduct user surveys. The current work on the multiprotocol interface is taking place on a wiki page unavailable to the general public. The Kadu project's forum is used to get feedback from the users on finalized mockups and prototypes.

So far, the work on Kadu Usability Project has been taking place primarily in Polish (it is safe to assume 99% of current Kadu's userbase speaks Polish and all the contributors are Polish) but essential documents have been translated to English (see: http://www.kadu.net/w/Usability) and Kadu itself has a working English translation.

The work on the project has started in July 2008 from conducting a survey with the developers of the project. I've used it to gather information about the goals of the project, intended userbase and usage conditions. The results are available in English here.

The next step was conducting a user survey. It's been conducted over the Internet in Polish (using LimeSurvey) and advertised on various Polish websites and forums about FLOSS/Linux. The survey has been active on 1 - 25.08.2008. During that time it has been completed by 5035 participants. It had two versions: one meant for people using other IMs than Kadu (including past Kadu users) and the other meant for current Kadu users. The version for the current userbase has been completed by 3018 participants. The goal of the survey was to obtain the demographic data about Kadu's current userbase including age, sex, computer skills, the operating system / window manager used, the usage of Kadu and its optional modules, other IMs and communication protocols used as well as gathering suggestions about Kadu's further development and introducing multiple protocols in Kadu. Because of the tremendous amount of data collected (the survey had a few open questions that were filled in by 1000 participants each on average) the analysis of the results has taken me two months and ended on 24.10.2008 with the publication of 35-pages long report. For obvious reasons, it hasn't been translated to English.

Basing on the data acquired in the developers and user surveys the personas for the Kadu IM has been built. Our primary persona - Piotr, represents the typical user of Kadu as far as demographic data goes (basing on the user survey). His usage of multiple communication protocols is more extensive than is typical for current Kadu users to help us focus on extending Kadu as a versatile multiprotocol IM.

The secondary personas - Andrzej and Kasia - represent two extremities that has been actually recorded in the user survey (albeit in lesser quantities) with all the other user survey paticipants falling somewhere in between them. Andrzej is a very experienced user of computers and a professional sysadmin. Because Kadu is primarily a Linux/BSD application its userbase tends to be more advanced than that of similar programs on other platforms. With Kadu becoming a Jabber client in the near future, the amount of extremely experienced computer users using Kadu may become even bigger - Jabber is a protocol of choice of dedicated and advanced users of FLOSS.

At the same time the userbase is certainly not dominated by such people because of Kadu's graphical user interface, ease of use and the word-of-mouth publicity it gets of a simple to get used to clone of a very popular Windows application - the official Gadu-Gadu client. For these reasons we can presume Kadu is an IM of choice for many novice Linux users from Poland. Kasia - a teenage girl with no particular interest in computers, is a novice user persona.

Supporting personas aren't based on actual research. They represent future users of Kadu that may appear when Kadu becomes a multiprotocol messenger, therefore becoming attractive for non-Polish people (in case of Albert, we've assumed that a new IM with no international userbase will attract dedicated and more experienced FLOSS users first) and after fully-functional, supported and mature Windows and OS X ports are complete (in case of Paweł and Kuba). Additionally from the mutliprotocol interface design standpoint Paweł and Kuba represent Gadu-Gadu only users we cannot forget about.

Our personas could use a bit more work, the scope of which is detailed in this document.

In 11.2008 the actual work on the design of the multiprotocol user interface has been started. The progress is being detailed on this page.

First the elements of the interface that need to be changed were determined. A review of other multiprotocol messengers has been conducted (with emphasis on Kopete, Pidgin, Psi and Gajim - the direct competitors of Kadu). Some work on use scenarios has been done and an effort to establish a concise dictionary of terms introduced to Kadu along with the multiprotocol/multi-account/metacontact functionality has been made. Mockups for most of the windows that need to be introduced or redesigned are ready though up for changes (so far we've been using Dia and QtDesigner to create mockups, right now I'm considering a combination of Flash and QtDesigner screenshots to create a clickable prototype). The first window we've designed - Account Manager - has been introduced into a recent test build of Kadu and is being tested by the user community.

Unfortunately, because of the amount of work this project requires to provide a top-notch user experience to Kadu users and our limited resources (we're students working on this project in our limited spare time for free) we haven't been able to fully conform to all standards of the User Centered Design.

I think the personas we've built should be used more extensively in the design process, more use scenarios should be created and confronted with the current mockups and the mockups should be coordinated to provide a complete specification of the redesign to the developers (that's what I hope to start working on more extensively in May).

Clickable prototypes (either built using Flash or other means or test builds of Kadu) could be used for remote or local user testing. An afore-mentioned dictionary of terms we're introducing into Kadu needs to be established and polished both in English and Polish.

On the other hand Kadu's configuration manager is in need of a redesign (thanks to its XML-based structure the redesign wouldn't be very difficult to implement for the developers) that would take into consideration the changes happening in the rest of the interface.

Therefore the input we'd like to get from the Season of Usability student includes:

• - polishing our personas and providing use scenarios to confront with our multiprotocol interface mockups to assure their functional completeness
• - providing heuristic evaluation of our mockups
• - working on establishing a dictionary of terms used in the redesigned interface (in English, and if possible, in Polish)
• - working with us on creating new mockups and prototypes that would take into account the aforementioned suggestions
• - usability testing and / or collecting feedback from users about the mockups/prototypes
• - providing heuristic evaluation of the areas of Kadu that are not covered by the current redesign

If there is enough time left we'll move on to working on redesiging Kadu's configuration manager.