Open Source Projects and Usability Professionals?

accountant asks: "A very close friend of mine is a qualified usability tester. She gained her Master's Degree last September and now has a good job with a mega corporation. Unfortunately, times being what they are that mega corporation is looking at a mega merger and mega job cuts. It's tough finding jobs in usability at the moment (a lot of companies haven't grasped the concept of usability, nor usability specialists, and a lot of dot coms have no cash to take on more staff). It's especially tough if you have less than one year of experience. Usability, especially on the desk top, is a major priority for Open Sourced environments like KDE and Gnome. Who's doing the testing in such Open Source environments? Are they doing sizeable studies, running focus groups and doing statistical analysis on user test results? Money is handy, of course, but there must be opportunities to feed a resume and help the Open Source community. Your thoughts and URLs please?"

No, but they should

[selectspec]

Usability testing tends to be primarily focused on GUI desktop applications. Right there you pretty much nix most open src projects. However, there are quite a few open source desktop apps. Mozilla for example doesn't do any formal usability testing, although the a good majority of the bugs submitted are "enhancements" of the usability nature. I'd imagine that during Star Office's proprietary days under Sun and StarDivision, that some sort of usability testing was performed. However, I doubt that any is currently happening for Open Office. The problem is that focus groups and usability tests cost money. I wonder if groups like Ximian are doing this sort of thing?

Sun

[Wesley Felter]

IIRC, Sun has a group doing usability tests on OpenOffice and GNOME. It will be interesting to see if the developers pay any attention to the results.

Not really the Open Source way

[Anonymous Coward]

Who's doing the testing in such Open Source environments? Are they doing sizeable studies, running focus groups and doing statistical analysis on user test results?

Frankly, that's not really the "Open Source" way of doing things.

Chapter 4 of The Cathedral and the Bazaar [tuxedo.org] describes this. "Release early. Release often. And listen to your customers." Open Source projects tend not to have one monolithic usablity test with focus groups, statistics, PHB's to evaluate the whole thing, etc. Rather, they have a constant stream of incremental releases, each improved by continuous user feedback. Sure, you could do an elaborate usability test with all the trimmings on a particular release, but by the time you got done with it, your tested release would probably be 2 or 3 versions behind the current release and the user interface may already have changed so much that your test results may hardly apply any more :)

That having been said, have you seen the Gnome usability project [gnome.org] (join [gnome.org]).

research the openreference interface

[Dostoyevsky]

I'm working on a project I call openreference [openpolitics.com]. It's basically object-oriented hypertext (sort of like Microsoft's SmartTags idea, but actually open).

Anyway, one of issues I haven't settled is user interface. I imagine openreferences as a sort of graphical version of the Unix pipe. The Windows Send-To menu is one possibile interface and I use it in my examples but I'm not sure it's the best interface.

If you want to find out more about my idea, read my proposal [openpolitics.com].

Re:Not really the Open Source way

[dannyspanner]

"Release early. Release often. And listen to your customers."

This assumes that your customers express their dissatisfaction, or even realise that something is wrong. Your average (non-geek) user will just accept that things are the way they are. How do you explain the dominance of Windows (BSOD et al) otherwise?

Re:Not really the Open Source way

[ebbe11]

Sure, you could do an elaborate usability test with all the trimmings on a particular release

Okay, this link [useit.com] describes usability tests on web-sites but but the basic principle outlined here would also apply to applications - and it would tie nicely into the "release often" paradigm.

Why not volunteer?

[winchester]

Why doesn't your friend volunteer to help in the gnome/kde/... project? I am sure they are more than happy to get your friend's input on usability issues.

Anyone can contribute to open source, that is the beauty of it. I would like to see an open source marketing group, to market open source software. An open source documentation group, providing the best documentation possible. An open source test group, to test scalability, reliability et. al.

Would be good for open source, to run a little more like a business. After all, this is a business world.

Usability, focus groups, and usability jobs

[jefraskin]

It is difficult, and perhaps impossible, to do excellent UI design with a distributed contributing community. Having an open-source UI is fine, but for an interface to even aspire to a high degree of usability, it must be (in certain ways) consistent. It must allow the formation of habits so that it can be used without conscious attention, and it must not cause the formation of habits that will foster errors.

To meet these goals, and there are of course others, requires very careful design by people who deeply understand the principles of cognetics. These people are uncommon even in the interface design community, and exceedingly rare in the programming community. The problem is compounded by the legions of the ignorant whose criteria go no farther than "I know what I like" or "This works well for me."

What can profitably be open is the design specification. Input from many sources can be of great value, so long as there is a moderator with the requisite knowlege. This is not to say that different moderators might might not make different decisions, but that the final design decisions have to be made by, at most, a small and technically knowledgable group.

A number of commentators have mentioned "focus group" testing. Usability is not tested in focus groups. Desirability is closer to the kind of quality tested by focus groups.

One way to find out what jobs are available is to join a professional group such as the ACMs SigCHI. Our local chapter (BayCHI, which meets in Palo Alto, California) distributes a list of current openings.

[e-mail]

It should be done free by students

[neves]

I agree. Usability teachers should give their students a open source application to test. Why spend time testing something you can't influence? Testing Gnome or KDE they would have chance to see their suggestions approved by the developers. Which other OS could be influenced this way? Is there any site that mantains a pool of open-source projects that want to be reviewd?