Monday Jun 09, 2008

How did our JavaOne talk go?

JavaOne 2008 is done...and there are lots of good memories. Jindra and I spent the week learning and practicing our talk, as well as wandering the pavillion and attending sessions. Our slides can be accessed here on the JavaOne Online site (note that you will need to log in as an Sun Developer Network member to view the slides -- registration is free and you won't be spammed so go for it). Also check out the slides for the other User Experience related sessions (TS-6929 Creating a Compelling User Experience, TS-6470 The Layperson’s Guide to Building a Better User Experience, and TS-5500 The Desktop Java Technology Lovers Survival Guide)

Speaking at JavaOne is totally exhilirating. Our session took place in a large room and was well attended. This year's JavaOne enabled folks to "pre-register" for each talk for guaranteed admittance, so we watched the number of prospective attendees grow during the week. In the end, about 600 folks signed up and 500 actually showed up. Neither of us has presented to this large of an audience before (we were really excited last year when our BOF had around 150 people). Since we were both new to delivering a technical session, we went through the material much more quickly than during our rehearsals. Next time we will be better prepared by making sure we have some extra material in case we finish too early -- it's easier to keep talking about stuff than to make things up.

Jeff and Jindra during Q&A

Here is a photo of us during the question and answer period of the talk. We definitely have a bit of that "deer in the headlights" look. One disadvantage of having your talk scheduled on Friday of JavaOne week is that you are nervous with anticipation for pretty much the entire conference. There are certainly benefits to presenting early and getting it "out of the way".

JavaOne attendees have high expectations, and since this year's conference featured a set of good user experience sessions, we were in very good company. Our audience was very supportive and we didn't lose many people after we started talking... If anything, I believe the folks wanted more details than our 101 level talk provided. Once we've reviewed the feedback on our audience response cards, we'll start planning for next year's session and take their comments in to account.

If you have some ideas for user experience topics that would be of interest to the wide range of Java developers who attend JavaOne, we invite you to leave your comments on this blog.

Jeff Hoffman is the lead user experience designer for Java Standard Edition.

Jindra Dinga devotes his time to improving the deployment experience of Java for both developers and end users.

Thursday May 08, 2008

JavaOne - the Pavilion

Today I was wandering around the JavaOne pavilion and took pictures of booths that showed stuff xDesign team has been working on. So enjoy! BTW: the pictures are randomly ordered ;-)

Because Sun is not the only one exhibitor in the pavilion, I also took some pictures of booths of other companies.

by Jindra

Wednesday May 07, 2008

A Quick Summary of One User Experience Talk at JavaOne

Jeff Hoffman is the lead user experience designer for Java Standard Edition.

It's the second day of JavaOne, and I attended a talk by Ben Galbraith titled Creating a Compelling User Experience. The talk was very well attended and almost filled the 826 person room (and it happens to be the same room that Jindra and I will be presenting in on Friday).

Ben is an entertaining speaker and put together a very slick presentation and demo. He conveyed key items that a developer needs to consider when designing a compelling user experience for their app. The presentation was peppered with quotes from the greats of usability and user experience design, including Alan Cooper, Jef Raskin, Donald Norman and Jakob Nielsen. A few of the points he made stick in my memory, so I'll share them with you:

  • Understand your user, and their expectations
  • Don't let your end user literally design your UI -- base your design on the goals they are trying to achieve
  • Get a visual designer to work with you -- UI design can be likened to fashion design, and you want your app to be "in"
  • Make sure your app is responsive, if the user has to wait more than a second for a response, their mind starts to wander
  • Respect the user's data, that is, don't lose anything the user enters in to the app

He also made some very positive comments about the new Java browser plugin that is included with Java 6 Update 10, and JSR-296 -- the Swing Application Framework. These features enable Java developers to create more responsive applications both in the browser and on the desktop.

Stay tuned... More photos from the show will be posted soon!

Tuesday May 06, 2008

Java One 2008 - Day One

As usual, JavaOne conference starts the second day in the week. The pictures below show a decoration you can see on streets, inside Moscone Center, as well as our friendly staff working at the registration desk.

This day started with James Gosling's talk. Right after him Rich Green came on the stage with his Java + You talk in which he mentioned how important is community and collaboration. During his talk he also showed us a demo of a application written in JavaFX. In addition to this day, the pavilion was open for the first time. On our pod, we showed some cool stuff our team is working as well as gave away some gifts.

by Jindra

Wednesday Apr 23, 2008

Speaking about User Experience at JavaOne

Jeff Hoffman is the lead user experience designer for Java Standard Edition.

Jindra Dinga devotes his time to improving the deployment experience of Java for both developers and end users.

Jeff and Duke at JavaOne 2007In a couple of weeks, my colleague Jindra and I will be presenting our process for creating a graphical user interface to the developers at JavaOne. In my last entry, I mentioned a set of user experience talks happening at this year's conference. Now I'd like to describe a bit more about how we developed our session and what's in it.

At last year's JavaOne, our merry little band of Java UE designers presented a very basic overview of user experience design best practices at a 9pm BOF. We dutifully put together a presentation with slides covering a variety of things, and cheerfully presented them to the much larger than we expected crowd. We were terribly nervous, but overall the experience was great and the questions were great too. Some months later the survey results came in and they weren't bad, but not great... Most of the comments were asking for more detail and more examples, so we started discussion about this year's presentation with that idea.

Based on the feedback, this year we are going to take a real example and walk through our process with it. Since we only have 50 minutes (and some of that time needs to be available for questions), we will try our best to reach the level of detail our audience desires. JavaOne Speaker

At the beginning of our presentation we will talk about why it is hard to create good GUIs and how important it is to understand the user's tasks and goals. Later on we take the existing command line process for configuring a network interface connection in Solaris (see Project Brussels) and make it over in to a GUI.

Jindra and I have spent years in the user experience field and we know that it's hard to follow an exact process for every project. We also know that making sure our designs work for our customers requires that we adhere to the principles of design, and we want to make sure that the developers out there understand how these principles apply to a real design problem.

If you're planning to be at JavaOne, sign up to come to our session (TS-4968). Also, if you'd like to say hello to some of the contributors to this blog, stop by the User Experience pod near the Spin-the-Wheel Game in Sun Booth at the JavaOne Pavilion.

Thursday Apr 17, 2008

Data Driven Personas

A week ago, I was in Florence Italy, giving a talk on a paper that a colleague and I wrote. The paper was a CHI Note, which described a new way to create "personas": fictional characters who represent user groups who will use the products that we develop. It's a lot like creating user profiles, but personas have names, photos, and other details added to them, so that designers and developers can meet the needs of, say, "Sarah" rather than a nameless, faceless "system administrator".

The process for creating personas has been around for about ten years, and usually starts by brainstorming user characteristics. Ideally, the details of the personas are drawn from facts that have been gathered through user contact. Sometimes, once the personas have been created, they are then validated, by conducting a survey or focus groups or by communicating with representative users in some way. Unfortunately, there are several drawbacks to the traditional method, usually revolving around trust in the accuracy of the personas and acceptance by the team that is supposed to use them. But it's a numbers game, too: if you interview 10 or 15 or 20 users, how can you ensure that they are truly representative of the larger population?

My partner in this paper, Nalini Kotamraju, is a user researcher at Sun, and it was she who initially proposed that we turn the traditional method around: rather than interview users first, and then validate the personas through a survey, why not conduct the survey first, and then interview them? Brilliant! Not only would the personas be an artifact of the data, but we'd have statistically significant numbers of users to base them on! Then, when we conducted the user interviews to add details to the personas, they would be "the right" (read that as "representative") 20 or 30 users, because we'd have done the stats first.

Our new persona creation process went well, but there were some surprises along the way... perfect for a paper submission :) Fast forward a little more than a year, and there I was in lovely Firenze, sharing our work with a receptive group. My favorite part was at the end, after the talk was over, when I was able to share more in-depth details with friends and colleagues. I even got a question and compliment from John Pruitt! But there were a lot of friends and colleagues who weren't there, so if you're one of them, I'll be giving another talk in May :)

Jen McGinn is an interaction designer in xDesign who is working to improve the user experience with software registration and Solaris system administration. She has an MS in Human Factors in Information Design and works out of Sun's campus in Massachusetts.

Thursday Apr 10, 2008

Calum's CHI Photoblog - Part 4


(My cold wasn't too bad today, as it turned out... throat much better, somewhat bunged-up otherwise but better that than having a runny nose all day, IMHO...)

Started the day at the Character Development session where Jen was presenting her data-driven persona talk—good to meet her in person at last! Also bumped into a few old (and a couple of new) faces there, including Robin Jeffries, now at Google, and Matthias Müller-Prove, with whom I co-authored a paper and presentation for CHI in Vienna.

Unfortunately I was paying so much attention to Jen's presenation that I forgot to take a photo, so you'll have to make do with this out-of-focus, mostly-people's-backs effort of her (left) chatting to Andrea Knight from Google, Christoph Noack from Bosch, and Robin (mostly obscured by Christoph):

Jen McGinn (Sun), Andrea Knight (Google), Christoph Noack (Bosch), Robin Jeffries (Google, obscured)

Another of the talks, on identifying personas using latent semantic analysis (essentially looking for similarities in the textual patterns of questionnaires completed by the subjects), was interesting too, although some of the resulting groupings presented in their example seemed a little dubious. Probably another one I ought to go and read the paper for...

I ducked out of the session after Jen's presentation to see Evangeline Haughney's talk on how she's used comic strips at Adobe to present user interview findings. This is obviously something we've done some work on at Sun as well, so it was interesting to see how her approach differed from ours. The most obvious difference to me was that she didn't use the traditional chronological comic strip format, but an information-rich montage of comic strip elements on each page. And also that she delivered it to the stakeholders as a printed comic book, rather than just in an email or on a website. (I really want to try this on one of my projects soon!)

Next up was the Collaboration and Cooperation session, which included a couple of talks on "co-located collaborative web searching" (aka "one person Googling while one or more others look over their shoulder trying to help") . One paper described a novel solution where the 'observers' who aren't sitting at the keyboard can contribute by suggesting search terms to the 'driver' via their bluetooth phone, on which they can also browse results pages independently of the other participants.

I went for lunch with Christoph Noack from Bosch, who's part of the OpenOffice user experience team, and Mike Terry, Ed Lank and Christine Szentgyorgyi from the University of Waterloo in Ontario—Mike gave the InGimp talk I attended on Tuesday. We chatted about his plans for analysis on the data collected via InGimp, whether the instrumentation might be abstractable to GNOME applications (since GIMP and GNOME are both built on gtk+), and about open source usability in general.

(Having spent some time trying to remember where he'd heard my name before, Mike suddenly pulled out his laptop halfway through lunch to show me a paper he was writing that referenced one of mine!)

Got back just in time to hear a case study from Intuit on the advantages (and pitfalls) of using users' real-life data, such as bank and tax details, in their usability tests. They cited a couple of examples where features that had been readily discovered by participants in 'fake data' tests were missed when users worked with their real data instead. To counter that, they also gave the example of the user associating a memory-jogging phrase with certain data to aid recollection months or years down the road—it was impossible to test the effectiveness of that feature, because the user naturally had to choose the phrase just before the test.

So to Bill Buxton's closing plenary, where he spoke about "not what can we do, but what should we do" with the expertise that the CHI community has. He showed how how great design borrows from and extends the great designs of the past, and how a product's development needs to include a consideration of the social and cultural implications of its success, rather than evaluating it in isolation. A suitably thought-provoking finale.

Bill Buxton

It's been a busy week, with some interesting sessions and some not so interesting ones, as always. Was good to meet up with the Sun folks, who I don't see in person very often, and I was particularly glad to meet some other proponents of open source usability—there still aren't many of us around! It's probably fair to say the only real disappointment this week has been the weather...

And with that, I'm back off to Dublin late tomorrow afternoon to catch the last few hours of my wife's birthday, so I guess I'd better try to find her something nice here in the morning :)

Wednesday Apr 09, 2008

Calum's CHI Photoblog - Part 3


Woke up this morning with a very sore throat and all the other tell-tale signs of an impending nasty cold. Still made it to my full compliment of sessions today, but might struggle a bit tomorrow... hope not though as at the very least I want to go along to Jen's talk!

Started today by attending the "25th birthday party" of The Psychology of Human Computer Interaction, the seminal book by Stuart Card, Thomas Moran and Allen Newell that more or less coined HCI as a phrase. Stu and Tom were on hand to talk about how the book came about (Allen having passed away in 1992), and a panel discussion led by Bonnie John reflected on the progress we've made since then with the likes of Soar, ACT-R, and various GOMS derivatives.

Stuart Card

My other session before lunch was Am I Safe?, where three papers were presented. The first looked at a novel way of visualising firewall alerts (traditionally jargon-heavy popups presented by applications like Zone Alarm), which amongst other things drew a connecting line between the window of the process generating the alert to a world map showing the location and details of the server generating the request. A novel idea, although the visualisation still looked over-complex to me and could use some work.

The next speaker presented a system that drew graffiti on the user's desktop when it determined that one of their installed applications needed updating to patch a security vulnerability—the more prominent the graffiti, the more serious the vulnerability. Hopefully, we wouldn't have quite as much use for that on Solaris as the Windows users in the study seemed to have :)

Finally, we heard from a team who investigated the effectiveness of the "phishing" alerts generated by IE7 and Firefox 2. Solaris users will be glad to know that none of the Firefox users in the study were fooled into giving their details to a phishing site set up for the study—many of the IE users did though!

I went poster-perusing again at lunchtime (bumping into former Sun colleague Nancy Frishberg in the process). Here are a couple that caught my eye today—click for the legible version:

Poster-3 Poster-2 Poster

In the afternoon, I went to:

  • Branding the Feel: Applying Standards to Enable a Uniform User Experience: A panel from Adobe, Microsoft, Google and SAP talked about the process of writing user interface guidelines, and providing supporting materials, in the context of a corporation for whom their user interface is part of their brand. Thought I might come out of it with some insights that might be useful in the many branding discussions we have at Sun, but in fact I came away with more ideas about refreshing the GNOME Human Interface Guidelines (an exercise which is long overdue at this point!)
  • Menu and Command Selection: First, a chap from Autodesk presented their "PieCursor" concept—a new take on tracking pie menus in which the cursor is the menu (or more specifically, the tool palette, to which it's really more suited). Their results showed it to be more effective than using a toolbar, but little mention was made of the fact that they also seemed to suggest it was less effective than their existing pie menu. (Unfortunately I didn't get a chance to ask why not...)

    Next up was a presentation from a Chinese team about their "tilt menu", in which the user summons a pie menu by tilting the pen or stylus (or it may be shown automatically, if the context demands). They select from the menu by tilting the pen further in the direction of the appropriate menu item. It looked like it might be a little awkward to use, to me, and as you might intuitively expect, their study showed that some segments around the pie were easier to select than others.

    Adaptive activation area menus were next, attempting to address the 'steering problem' of cascaded menus by manipulating the 'hot zone' for accessing the submenu once the pointer landed on a parent menu. I could be wrong, but the basic method described (two other variations were also shown) was awfully similar to what I thought GNOME already did, but I wasn't sure enough to raise that point at the time :)

    Finally, an algorithm was presented for automatically improving the layout of hierarchical menus based on decision time, item category and frequency of use. The speaker was pushed for time, and was not presenting in his native language, so it was not entirely clear to me whether this was intended to be used during the design phase, or on the fly on a per-user basis in the real product. Think I'll have to go and read that paper tonight...

Tuesday Apr 08, 2008

Calum's CHI Photoblog - Part 2


Okay, slight confession—I skipped the first session this morning to go shopping for some essentials that I forgot to pack! I did take a few snaps on the way though, and since I didn't take any other conference-related photos today, I at least ought to show you some of those instead.

Copy of Michelangelo's David, and Hercules and Cacus, Piazza della Signoria

Ponte Vecchio

Santa Maria Novella

So to the sessions I did attend:

  • Beyond end-user programming. I particularly wanted to go to this session because of the InGimp talk, to hear how Michael Terry et al. instrumented the open source GIMP image editor (available on a Solaris desktop near you) in such a way that all the information collected is publicly available. Since the GIMP uses the same widget toolkit (gtk) as the rest of the GNOME desktop, I was interested to hear how easy it might be to abstract this technique to instrument any GNOME application, perhaps even making it a standard feature (for development releases at least). I did specifically ask this question, but received a fairly non-committal response :) I'd still like to follow up on this later, though.
  • Meta-CHI was part of this year's alt.chi programme (a track causing some controversy in itself), and the three talks here concerned how to improve the way we do 'interaction criticism', a re-appraisal of the three classic principles of 'early focus on users and tasks, empirical measurement, and iterative design', and a comparison of the way that quality criteria are applied in the fields of science and design. As with the opening plenary, all a bit introspective and academic for my tastes, but at least the clue was in the name this time so I could go prepared!
  • Friends, Foes and Family. Social networking is a big theme at this year's conference, and also concerned a couple of the papers in this session. Have to say I was disappointed overall, very little made me think "oh, I never thought of that."

    The first talk was about how people assess 'attractiveness' on online dating profiles (conclusion: it's mostly about the photo—no surprise there), but without any real explanation as to how the research might be useful. "Maintaining friendships after a residential move" had the most potential out of this session, but again there were no real surprises; the one interesting snippet perhaps being that decreasing email frequency after a move does seem to have a negative impact on perceived 'quality of friendship', but that increasing the frequency doesn't have a positive impact. The other two shorter talks were about whether friends on social networking sites are more persuasive in spreading content than casual acquaintances or 'foes', and how and why parents tend to record personal information on their online calendars at work, and the sort of problems this can create. The latter (a Microsoft Research project) was particularly surprise-free, and seemed to me to highlight shortcomings in Outlook's calendaring capabilities compared to other online calendars as much as anything :)

Calum's CHI Photoblog - Part 1

Calum Benson is a usability engineer in the desktop team in Dublin, Ireland. He has worked at Sun for about eight years, primarily contributing to the open source GNOME Desktop project's usability efforts, to the usability aspects of Sun's integration of that project into Solaris, and currently, to the design of various GNOME-based applications for OpenSolaris.

Maya asked me if I'd file a few blog entries while I was at CHI 2008 in Florence... so here goes.


I arrived at my hotel around lunchtime, but although I was only coming from Ireland, I'd been up since 3.30am to catch the first leg of my flight to Amsterdam. So I was too zonked to do much other than wander up to the conference centre to register:

fortezza de basso

sculpture outside fortezza

sunlight canopy

tree in bloom

welcome to chi


Everything kicked off for real at 8.30am in something resembling an aircraft hangar, with the opening plenary given from my compatriot, Irene McAra-McWilliam, from the Glasgow School of Art—perhaps best-known outside the UK for being housed in Charles Rennie Mackintosh's finest building. Irene talked about the evolution of design, structuring her talk around a common feature of Florentine architecture: the Rose Window. Personally I did find it a bit of an introspective talk for an opener, perhaps because I don't have any formal training in design so I couldn't immediately connect with the content. But given the theme of this year's conference ("Art Science Balance"), that was perhaps partially the point...

Opening plenary

The sessions I attended today were:

  • Interactive Image Search, where Cue Flik, a Microsoft research project, was demonstrated. Here, the user could identify image search results that were most like the ones they'd been looking for (the example used was a 'product shot', which often shows a product on a white background), to not only refine the results but to apply to future searches. Thus, in this example, you could apply the 'product shot' rule when searching for anything from shoes to suitcases, and expect to find the top results showing those items on a white background.
  • Usability Evaluation Considered Harmful? Bill Buxton and Saul Greenberg presented their paper highlighting the dangers of blindly applying usability evaluations, to some extent because academia has come to demand them, sometimes at the expense of rejecting research into more challenging areas of usability and/or resulting in the mis-application of "weak science". Followed by a panel discussion (which naturally concluded that both usability evaluations and non-empirical methods have their place, and that the key is in choosing appropriately for the problem at hand).
  • Improved video navigation and capture. Three projects were discussed here, the first being research into automatically capturing meetings for live or delayed webcasting in a more engaging fashion, with appropriate cuts, tracking, close-ups etc. depending on who's talking—could tie in nicely with some of the telepresence projects at Sun Labs :) The other two projects, DimP and DRAGON (demos available by following the links) both provided the user a means to navigate through video by directly manipulating the objects in the frame, rather than the traditional means of dragging a slider. Both showed useful productivity gains for certain types of task, although I have to admit they weren't the sort of tasks I'd necessarily expect most people to encounter in their daily interactions with videos.

Inbetween sessions, I had a wander round the posters (which, this year, are changing every day):


In the evening, we had the official opening ceremony, with a plethora of Tuscan wine and edibles...

Table decoration

a string quartet...

String quartet

... and a few exhibits to play with! This is Snap and Grab, a video wall that you can interact with from your cellphone:

Snap and Grab

And this is Remote Impact from DistanceLab, in which you can punch (and be punched by) a remote opponent!

Remote Impact

Monday Apr 07, 2008

User Experience Blooms at JavaOne 2008

Jeff HoffmanJeff Hoffman is the lead user experience designer for Java Standard Edition. He's been designing both consumer and developer oriented products since before the boom.

Springtime is in the air! The flowers are blooming and the trees are sprouting branches. Along with nature's beauty all around us, it's a sign that we're getting closer to a special time of year for Java developers...JavaOne!

I've been attending this show since 1999 and have always been amazed at the array of Java development related topics -- except for putting good user experience design practices to work for your applications. Sure, there are always wonderful discussions about adding fun visual effects like animations, reflections and transparency to help make your application stand out from the crowd. But developers have deeper questions -- like how to make an application work the way their users expect.

JavaOne SpeakerI'm very excited to report that the 2008 JavaOne show (May 6-9 at Moscone Center, San Francisco) will include at least four technical sessions focused on user experience. There are also a whole bunch of sessions on the latest fun stuff you can do with Java 2D and 3D Graphics, Swing, JavaFX Script and more. I'll be holding a session along with my xDesign colleague Jindra Dinga called Designing GUIs 101 (TS-4968) where we will use an end-to-end example to illustrate our straight forward, repeatable process for designing GUIs that meet the user's needs effectively and efficiently.

Our talk will cover the important phases of a good design process:

  1. Discovering the User's Goals and Tasks
  2. Gathering Requirements
  3. Defining the Task Flow
  4. Designing the GUI
  5. Gathering Feedback

And how we apply these practices when we are designing a specific feature (based on a current command line interface) within a GUI desktop tool. I'll be writing more about our technical session in the weeks right before JavaOne.

To round out this "mini-track" of User Experience sessions, I've discovered three talks from folks who (like me) are interested in enhancing the user's experience with software and are spreading the word about how to make sure it is properly considered during development. They are:

Looking forward to seeing you at the show!

Monday Oct 08, 2007

User Research at Innovation@Sun

Nalini Kotamraju is a user researcher in xDesign, and a PhD in Sociology. She has a penchant for research methods and telling it exactly like it is.

Jen McGinn and I recently had the honor of giving a talk about user research at Innovation@Sun, a gathering of Sun's top engineering talent. This illustrious group count among their ranks people who are pioneers in Java (of course), but also in computer graphics, routing security, cryptography, and large-scale distributed computing. Many great technical brains, many patents in pockets. An intimidating group, by most measures.

Jen was presenting (I was back-up) about user research that we had done last year for an organization in Sun. The research findings themselves are terrific and already being applied within Sun. What we wanted to share with this audience was the innovative way in which we conducted the research, and to remind the audience of the importance of understanding the people who are ultimately often the end-users of technical innovations.

One might imagine that such an audience, gathered to exchange information about advances and challenges in the realm of engineering might be -- at best, apathetic -- to a presentation about users and user research. Or at least I had imagined such a response from the audience. Instead, I found that many of the people who stopped by during the poster session or who asked a question after the talk were not only receptive, but were even enthusiastic about user research. In the formal settings, as well as over meals and in hallways, these engineers asked questions about how we think about understanding users and, more often than not, wanted to know how they could utilize user research in their own work for Sun.


xDesign is a software user experience design group at Sun.
Follow us on Twitter : Flickr : Blog (see feeds below)


« July 2016