It’s kind of ironic, not to mention ever-so-slightly depressing, that both the ‘rival’ UI testing suites du jour only mention Linux in their acronyms and descriptions, when a key part of the underlying technology on which they depend (the accessibility framework) was largely developed on Solaris boxes in the first place
Presumably the suites themselves should also run equally happily on said ancestral platform, so it would be nice to see a bit of a mention at least…
Category: GNOME
MS Office 12 UI blog
Jensen Harris from Microsoft has started a blog about the new Office 12 UI.
A decade of usability
Yikes, it’s ten years today since the first real project I worked on (for money) produced its first real deliverable.
I was employed by Logica (now LogicaCMG) in Cambridge, but was working with a team of other usability folk from Logica, Admiral Consulting (now part of LogicaCMG themselves), Microsoft, GUI Designers and others at Reuters’ now-defunct Usability Group just off Fleet Street in London.
With insufficient funding to build their own in-house team, Reuters’ Greg Garrison decided to form a virtual team’ of usability consultants from various companies, who could be called upon as they were required. The rest of the time, they would be back in their own offices, being paid by their own companies. (In practice, it didn’t quite work like that; most of us were there 10 or 12 hours a day, 5 days a week, blowing the budget on a variety of wild ideas.)
And the deliverable? A multimedia extravaganza of usability examples, guidelines and icon libraries on CD-ROM, all produced in Macromedia Director, Microsoft Visual Basic, and Adobe Acrobat and Photoshop (all of which were at about Version 3 in those days). And burned on the office’s own CD burner– which cost about three grand, and took half an hour to burn a disc.
I can’t say that the work was always a whole lot of fun… long hours, personality clashes, artistic differences and inter-company rivalry were never too far from the surface, and because the virtual team thing was considered innovative at the time, Greg was always trailing all sorts of media and BPR types around the office (and particularly the usability lab that was part of it) while you were trying to get your work done. But most of the stuff we did, and the CD-ROM itself, still stands up pretty well, I learned a whole lot from the people I was working with, and made some good friends. (The occasional spot of wild partying in London was a bonus, too…)
Add to Panel redux
A few weeks ago I posted my initial reaction to the new ‘Add to Panel’ dialog that had snuck into Ubuntu Breezy, and it wasn’t very positive
It also provoked the most comments I’ve ever had on a blog entry, many in agreement, and some not.
To his credit, Manu took my comments very well, pointed out one or two things about the design that I hadn’t noticed, and asked me if I’d blog about it again once the design had settled down a bit.
So, here I am looking at it again.

On the positive side, it now fits on an 800×600 screen, and the categories are all filled out a bit more, so the initial “what a waste of space” reaction is gone (although it still hits a bit if you scroll down to the Internet and Multimedia sections, which only have one item apiece in them.)
To be honest, I don’t think I’m ever really going to like the OSX SystemPrefs-like layout; the first impression you get when the window opens is that the icons are arranged almost randomly, because they’re sorted in three dimensions rather than the simple alphabetical list we’ve all become used to. The description of each applet is also rather hidden away (near the bottom of the dialog), compared to the old design where it’s right beside the applet in question. Tooltips might work better here.
The search box is promising, but needs a little more work to be spot on (ignoring, for the moment, the fact that it works like nothing else in GNOME). It filters the visible list based on what you type, searching both the applet names and descriptions. So, typing ‘sound’ will leave you with just the “Volume Control” applet on display. It doesn’t actually select any results, though, even when there’s only one matching applet, so you can’t just hit Enter straight away to add it. Nor are the results ranked in any way, so if I search for ‘log’, I get ‘Blog Entry Poster’ and ‘Log Out’ in that order, even though it’s most likely ‘Log Out’ that I’m looking for. That’s probably not a big deal with the small search domain we’re dealing with here, though.
Another problem is that the applet descriptions can’t (and shouldn’t) include every word that a user will search for. As an example, the first four search terms I used to try and ‘find’ the Rhythmbox applet were MP3, CD, player and jukebox, none of which gave me any matches. We’d probably need some sort of hidden keyword system for this to work as well as Apple’s does.
Other minor gripes are that it doesn’t currently allow multiple selections, and longer applet names get clipped. A bigger issue is the icons on display don’t seem to respect the current theme, which will need to be fixed to be considered accessible.
Verdict: it’s a lot better, and I expect I could live with it. But personally I still prefer the old one 
IE designer switches to Firefox
An interesting defection indeed.
Usability slip-up
So, I guess 2.12 is the first major GNOME release in a long time where we’ve really made no attempt at any sort of UI review. Either this means we don’t need them any more, which would be great, or we (the GUP team) just suck granite pebbles for not getting around to it this time 
Subjective behaviour
If you’ve been getting “I don’t think you meant to enter that as your title” messages from NewsBruiser/b.g.o recently, bolsh has the answer:
[18:31] bolsh There's something that is intended to prevent you from
setting your password as a blog title by accident
[18:31] bolsh But if you've never set your password, it always returns true
[18:31] bolsh Don't know why
[18:32] bolsh You should blog about it :)
So, just reset your password, and everything will be hunky dory.
What page in yonder window loads?
Actually, I occasionally find the IE feature to which Havoc refers vaguely useful… he’s right that it never opens the page you want to go to next, but particularly if it’s a portal, it might well contain a link to one that you do1, and it’s harmless enough if it doesn’t, provided the page load is quickly interruptible. Admittedly you could just Shift-Click the link in the original window to achieve the same effect, but I usually find shortcuts more convenient than modified clicks.
It also reinforces the fact that your browsing history has been carried forward from the previous window– a trick that Firefox would do well to learn.
1The Macintosh version of IE made portal browsing even easier by providing a Page Holder tab in its sidebar that could hold all the links on a page for you, so you didn’t have to keep navigating back and forward… would be nice to see something like that in some other browsers too.
Applet shock
Eew, just updated my Breezy box and was confronted with this unhappy sight when I tried to add a new applet to my panel:
Don’t do it, I beg you
I presume it’s early work in progress, but it always seems so much harder to find what I’m looking for in these 2-D categorised list things (especially with no typeahead working yet). And just look how much space it’s wasting… by default, it doesn’t even fit on an 800×600 screen, and more than half the window is empty.
So much for a quiet weekend…
Wow… haven’t seen my name bandied around these parts in a while, let alone by luminaries like Havoc and Jeff
Guess that means I ought to throw in my €0.02…
Should GNOME have a “design team”? Well, I don’t think there should be any elite group of folks locked in a room who design every window we see on our screens. I do think it would be cool to have a core group of experienced usability folks whose primary role was to act as consultants to individual projects, from start to finish, responsible for whatever task analysis, requirements gathering, design and usability testing is needed along the way. And then maintaining the HIG in their copious spare time
But that’s not too far from what the GUP originally set out to do, and the UI Improvement Project before it, and neither really gained any traction in that particular area. (Nor, to the best of my knowledge, have any other big open source projects had any more success– I’d love to hear about any that have.)
FWIW, I do think there’s sometimes a downside to having two of our most experienced usability folks work for the same company, because especially with the creative sort of work those guys do, it’s always tempting (and, frankly, a lot easier) to bounce ideas around and work together face-to-face, than to seek approval and feedback from the community at every step along the way. That’s not a personal criticism of Seth, Bryan or anyone else… at the end of the day, those of us who are lucky enough to be paid to work on open source software have a duty to our employers as well as to our community, and occasionally that means not being able to do things exactly how we’d like. Unfortunately that can lead to accusations of closed-ness from time to time.
(I’d also love to see the Novell usability folks participating more… we all know they do great work from their GUADEC presentation, but personally I miss their regular input elsewhere in the community.)
I wouldn’t go so far as to say that HCI work is as subjective as Joe says, either, although nobody could claim it’s a science. But there are certainly recognised methodologies for measuring usability, both quantitavely and qualitatively, provided you know your users. That’s where we still have a bit of an issue, I think, and despite Havoc and Jeff’s pleas in the past to develop personas for the GNOME project, we still haven’t got around to it.
Finally, as for Jeff’s “haranguing” me to lead the GUP… well yes, it’s true I don’t really have the time, and I doubt I’d live up to the expectations set by Seth anyway. But as Sun are moving more towards shipping less-customised versions of GNOME in OpenSolaris and as the basis of JDS, I certainly hope to be channelling a lot more of my energies directly upstream in the not too distant future. So maybe one day 