Getting to know you-- Usability Meets Open Source
CHI 2002 Workshop and SIG, Minneapolis USA, April 20th 2002
Calum Benson, Sun Microsystems Ireland
What is CHI?
CHI (Computer-Human Interaction) is the largest annual conference for usability folks, organised by the Association for Computing Machinery (ACM). This year was CHI's 20th anniversary, and it was attended by nearly 2000 delegates, including famous names such as Jakob Neilsen, Donald Norman, Jared Spool and Ben Shneiderman. And of course, yours truly :o)
What was the workshop about?
Free and open source software is gaining acceptance more than ever before. However, few open source projects incorporate usability into their development lifestyle to the same extent as modern commercial software. The workshop was organised as a forum for usability students and professionals to discuss their issues and experiences with usability in open source projects, and to find ways of promoting the notion of usability within communities that haven't traditionally practised it.
A special interest group (SIG) session was also organised for later in the conference. This was open to all CHI delegates, as opposed to the workshop for which pre-conference registration was required.
The workshop was co-organised by Nancy Frishberg, Calum Benson, Suzanna Smith and Andrea Mankoski (all of Sun Microsystems), Anna Dirks (Ximian Inc.), and Seth Nickell (GNOME Usability Project). Unfortunately, because CHI2002 happened very soon after GUADEC3 in Spain, not everybody's budget allowed them to attend both, so only Nancy and I were able to make it to Minneapolis.
What happened at the workshop?
Nine people attended on the day, including the two co-organisers. The fields represented were:
- government (US Navy) and quasi-government (European Patent Office)
- education (K12) (University of Missouri's Shadow project)
- university training in usability (UC Berkeley, School of Information Management and Systems)
- learning within the professions (SRA Key's work on Linux)
- large corporations with an interest in open software development (Netscape's Mozilla; Sun Microsystems' work on GNOME).
The organizers brought several topics for discussion. The participants were quite ready to share additional issues:
- open source software development as a potential example of computer-supported coopeative work (CSCW)
- extending our UI work practices; identify what won't extend to big audiences
- Human Interface Guidelines, standards, documentation
- Technologies used to get the work done: email, publishing, phone, IRC, face-to-face, learning and mentoring
- Design process and testing
- Notion of authors and authority
- "We are not our users"
- How to involve end users in open source community? What tools can they use for their contributions? (CVS inappropriate mechanism for usability work)
- How can usability specialists make sure their contributions are credibly accepted?
- Encourage OS community to embrace known HCI practices and techniques
- Communicate usability attributes that are built into opens source software through documentation (separate from "standards")
- Separate "building usable software" (valuing usability) from "dumbing down" in minds of open source developers
"The Open Source community" is not a single unified entity as the phase would seem to imply, but rather a lot of efforts that agree on general principles (e.g. Licensing). Attention to usability principles already present in GNOME, KDE, Mozilla,OO.org, Shadow
- Sharing a framework for cooperation between open source developers and HCI people (e.g. tools donated from large corporations)
- Improved methods for bringing usability methods into open source development (best practices + forum for sharing)
- Why should CHI people care about open source software?
- Job market 5 years from now
- Not just about hobby hacking any more
- Large corporations can be co-operators
- Open source software will "be there"; desktops and devices
- Potential for innovation greater than with proprietary software
What happened at the SIG?
As well as most of the attendees from the earlier workshop, we had about another 30 attendees, representing companies such as Unisys, KPM, IBM, and French Air Traffic Control; and educational institutions such as the Open University, Sheffield-Hallham, UCBerkeley, U Michigan, U Cape Town, and U Paris Sud. We also attracted a couple of big cheeses from SIGCHI who were interested in promoting open source usability in industry journals and at the SIGCHI committee level.
As we introduced ourselves, we posed questions and stated our hopes and priorities for being involved in activities related to usability and open source software development.
- How to insert usability into open source development process?
- How to improve feedback from product users to product developers?
- How to encourage student projects in usability testing, evaluation and design standards? A single student project about open source software is unlikely to make much contribution. Several dozen university student projects over the course of 2 or 3 terms coordinated to build on one another are more likely to provide a comprehensive evaluation of a particular open source project and point the way toward redesign efforts.
- Open source software is still at state commercial software was at in 1980's: which got better by whining
- Creation of tools for developing better interfaces, and incorporation of UI processes into product development
- Licensing issues are separate from development
- Usability is more than "testing"
- Best UI happens early + iteratively
- Do we have innovation in open source, or are we just replicating proprietary s/w with open licensing?
- Important for innovation to get in at start of development of new products
- How to communicate HCI results to open source developer community? Code? Samples? Documents?
- Who has the itch and who scratches it?
- We know of successful HCI design practices, need to fan those out to various open source projects
- Mozilla = 100,000 individuals downloading, 50,000 individuals submitting code. Most recent HCI achievement is the designation of a designer per component (starting with the Mail/News component)
- Need to devise HCI lifecycle practices that map onto existing open source development lifecycle
- Offer notion of patterns, process and cooperation as tools to open source community
- Design methods leading to "responsible development"
- Promote respect for end users
- Meritocracy = "doing"
- Many users, and learning, are two attractive motivators for open source developers
Topics that came up where GNOME is already ahead of the game:
- File "ease of use" bugs alongside functional ones, ensuring proper recognition as a category of bugs, with stratification (crashing, severe, etc.)
Create a "UserForge" (similar to GNOME Usability Project, but not specific to one open source project)
Open source is a great way of getting software localized (Cf. http://www.naci.org.za). E.g. South Africa recognizes that proprietary/commercial software companies may not be interested in localizing software in the many traditional or official languages its citizens use. Open Source software and standards holds promise for software development with appropriate localization.
Some of the actions that people took away from the SIG were:
- Announce the workshop and SIG activities to an "open source for librarians" group
- Focus student project(s) on open source development
- Bring open source i18n efforts in Open Source and Usability to the attention of the existing ACM i18n SIG
- Bring initiative to SIGCHI exec committee
- Think about how the group could coach and mentor open source developers in usability
What impact could the workshop and SIG have on GNOME?
Publicity, and more visibility of GNOME and other open-source projects in education, for example. Other papers are already being written that reference the work done on GNOME usability, and a write-up of this CHI workshop is expected in a future edition of Interactions, the CHI periodical. And no publicity is bad publicity, right...?
Some of the workshop and SIG attendees from the educational field have already vowed to start asking students to carry out usability evaluations on open source applications as part of their coursework, rather than just the usual MS Office apps. One example of such an evaluation can be found at http://www.sims.berkeley.edu/courses/is271/f01/projects/StarCalc.
- As IBM, Sun, HP, Dell, Apple and others continue to adopt open source software to complement their proprietary systems, they will be employing more and more of those students who have benefited from working on open source software as part of their studies. Therefore GNOME is likely to be in the fortunate position of having more professionally-qualified HCI people at its disposal who have a grounding in the open source ethic and development process.
Some questions GNOME needs to ask itself:
- How can GNOME encourage more feedback from non-technical users? IRC is an unfamiliar environment to many who are used to ICQ, AIM, MSN etc., and the list of GNOME mailing lists is long and mostly technically-oriented. Rarely a day goes by without some poor newbie's question being answered with "you need to ask this on list x, this one is for y". (Even just changing the name of gnome-list to something like gnome-users could help here).
How can we make the GNOME Usability Project work better? One of the original hopes (as with the UI HitSquad before it) was that it could act as a kind of UI "consultancy" for the community, an idea that was also suggested independently at the SIG. But so far we haven't quite pulled it off.