d e s i g n @ S u n

Jakub is a User Researcher in xDesign and works mainly on Java tools. He is pursuing a Ph.D. in Cognitive Psychology.

As a part of my doctoral studies I am teaching a class on Environmental Psychology at Charles University in Prague. This subject fascinates me with its focus on application in the real world. It is not very theoretical and has lots of applications in our everyday life.

Knowledge about humans from another psychological disciplines like cognitive or social psychology as well as from this discipline's own applied research is directly used in architecture, urbanism and environmental design for making environments more friendly to its inhabitants.* Great attention is paid to studying interactions between people and their environment in the real conditions which takes into account the whole complexity of real settings.

Environmental Psychology provides answers to questions such as: How should we plan cities so that people find them as a desired place to live? How should buildings be designed so that people are not getting lost in them? What recreational areas help people to recover from daily stress? What kinds of front gardens discourage burglars from invading the place? How should the lack of privacy in hospitals be compensated without major changes to their current operation? What kinds of maps are really helpful in wayfinding? What office arrangements facilitate productivity best? What classroom arrangement keeps pupils' attention focused on the topic? What personalities finds certain landscapes unforgettable?

Focusing on peoples' needs and using knowledge about them in designing for them rather than taking in account just aesthetics, architectural expertise and superficial economical aspects gradually becomes common practice in urban planning and architecture. This approach does not effect "only" peoples' satisfaction but proves to be highly cost effective. Focusing on design and analysis of people's needs slightly elevates costs in the initial phase, but increased productivity and many other benefits save money in the long run.

Does this sound familiar to you? Yes, there are many many similarities in designing living environments and designing user interfaces for software products.

Hungry for more information? Sign up for my class:) Charles University, Faculty of Arts, room n. 338, every other Thursday at 5pm.

*My class does not focus on ecological preservation as the name could suggest, but on interactions between human and their environment.

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.

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!

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.

Today, it was CommunityOne day here at JavaOne. CommunityOne day started with a keynote given by Ian Murdock, who was talking about how Sun participates in open and diverse technology ecosystem; he also invited Jonathan Schwartz on the stage.

Later on, there was a panel talking about pros and cons of various approaches to community building. At the end of the session, Rich Green announced the first release of a new binary distribution of the OpenSolaris, which is now available for free at www.opensolaris.com.

Rich Burridge is a member of the Accessibility Program Office at Sun Microsystems. In his 21+ years with the company he has had a variety of jobs and is currently helping to develop Orca, a free, open source scriptable screen reader for people who are blind or have low vision. He also has one of the most popular blogs at blogs.sun.com.

As you may know, the Accessibility Program Office at Sun is part of the xDesign group who have the blog you are reading now.

Two of my colleagues in the APO, Peter Korn and Will Walker have accessibility related blogs. I too occasionally post on accessibility related issues in my personal blog, specifically in the accessibility category.

All Sun products should be accessible to people with disabilities. The APO is here to help make sure that happens. It has it's own little spot on the web where you can find lots of useful accessibility related information.

Some suggested links to get you started are:

• Accessibility Quick Reference Guide

  All about accessibility: what it is, why it is important, tips, and additional resources.

• Towards Accessible HCI

  A broad overview of the software human-computer interaction (HCI) issues surrounding access to computer systems.

• W3C User Agent Accessibility Guidelines 1.0

  This document provides guidelines for designing user agents that lower barriers to Web accessibility for people with disabilities (visual, hearing, physical, cognitive, and neurological).

Jiri Mzourek is a senior manager in xDesign, responsible for Sun Developer Products and SOA/BI. In his spare time, he evangelizes usability in the Czech Republic by organizing SIGCHI meetings, World Usability Day, and working closely with the Czech Technical University on usability and accessibility related projects.

Exciting news!!!

Have you ever dreamed of developing an accessible Swing GUI the easy way? Stop dreaming and checkout the new NetBeans a11y plugin!

Let's start from the beginning:

The xDesign team in Prague has a long history of cooperation with our local Czech Technical University in Prague, Department of Computer Science and Engineering. For example, together we built the first usability lab in our region, organized a local World Usability Day, and ran Czech SIGCHI. The last 12 months we also cooperated on the a11y plugin for NetBeans. The main goal for this plugin was to allow Java developers without any special knowledge of accessibility to develop accessible Swing GUIs in NetBeans.

After a year we are proud to announce the availability of version 1.0 (for NetBeans 6.0 and 6.1)!

How you can use it:

1) If you don't have it already, get NetBeans 6.0 or 6.1 from www.netbeans.org.

2) Download the a11y plugin from a11y.netbeans.org , or use the AutoUpdate (Beta) functionality built into NetBeans. If you downloaded it manually, once you're running NetBeans, from the pull-down menus go to Tools -> Plugins, select the "Downloaded" tab, and click on the "Add Plugins" button; then browse your disk for the downloaded plugin (nbm file) and select it.

3) Now, start the NetBeans GUI Builder (for example add a new JFrame file into an existing Java project.)

4) Go to the Window menu and select "A11Y Result Window". That's it!

Now, when you edit the GUI it will automatically be checked for accessibility. The findings will be sorted into 3 categories (Errors, Warnings and Infos), which will be described and also will have a suggestions on how to fix them. By double clicking on the findings, you accept the suggested fix (for example, double clicking a "Name" error would add an appropriate accessible name). More details (for example checking of tab-traversal) is described in the documentation.

Enjoy!

Many thanks from all of us to the Sun External Research Office for their financial support, Tomas Pavek (NetBeans engineering manager) for technical support, Max Sauer and Martin Novak - who wore two hats during the process (Sun QE and CTU students) and for their contributions and cooperation with all of the other CTU students.

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.

In 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.

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.

A former colleague pointed me to using Wink freeware, to do capture screenshots and create screencasts. Click the picture to check out a quick demo I created using Wink. Here's the exported PDF of the same session.

Since you can read all the features on their site, here's what's valuable from a UI designer's point of view. I've used Wink as a quick way to capture existing interaction behavior and mock up new interaction behavior to show engineers. It also enables me to take many screenshots as I walk through an interface without interruption! I no longer have to pause after taking each screenshot to save and name it in another application such as Photoshop.

Pros

• Lightweight and fairly easy to use
• Captures the context menus and cursor (In My-Old-Screenshotting-Method on Windows, using the " Alt-Print Screen" dismisses the context menu before the screenshot is taken. You can only capture context menus if you capture the entire display with "Print Screen". And either mode does not capture the cursor)
• Many different types of output - Flash movie, PDF, HTML with a choice of saving images as pngs, bmps, jpg, and more
• FREE! (vs products like Camtasia)

Cons

• By default the graphics for callouts and buttons aren't the prettiest. You can create your own custom callouts, but afaik, cannot create custom buttons.
• Once you define the screenshot area, it's a fixed width and height. So anything that falls outside of that area is chopped off, e.g. dialogs or menus that extend past the area are cut off
• Doesn't remember user choices well and has some annoying defaults for where it saves files. I've accidentally overwritten files :(

Give it a try! Alas, it's only for Windows and some Linux platforms.

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.

Kartik Mithal is the director of the Sofware User Experience group at Sun.

I attended 2 days of the Mashup Camp in Mountain View. Registration was free and included breakfast and lunch on both days plus drinks and munchies after the days session on the first day. The camp was run as an ‘unconference’ which is designed to promote the equivalent of hallway conversations and networking. It was like 2 days of non-stop networking. These notes are a series of observations, and in that they are somewhat like an ‘unconference’.

Note their logo emphasizes the jigsaw puzzle aspect of mashups

There is a lot of energy in the mashup space. Lots of companies and lots of developers see this as a major source of applications going forward.

Company sponsors for Mashup Camp

Ages were an interesting mix; some younger folks, but lots of middleaged folks (mid 40s and above). I expected to find mostly young people, but was surprised to find older folks.

There was discussion about consumer masuhps, enterprise mashups vs consumer mashups, and desktop vs mobile mashups.

Map based mashups were the most popular, e.g. locating people, and telling people about the location they were in. Some driven by the location data available from cell towers, and some on cell phones with built in GPS’s.

Because mashups are ill defined, and because they are a hot topic, and because they are web based, they’ve become the catch all for anything to do with the web. For example, in one of the sessions we got into a long discussion about the javascript capabilities of different browsers on different phones.

Speed Geeking

Modeled on speed dating, mashup developers had 3 minutes to tell us what their mashup was. There were 20 mashups and we were split into 20 groups (mine had 2 people). My group started at table 4. The developers had 3 minutes to tell us what they’d built, then we went to the next table. The developers got better at their pitches, so after about the 3rd table, the pitch was much smoother. It all got mentally very intense, and after a while everything blurred.

The conference schedule was pulled together on the first day. Anyone who had a topic to discuss or had a question to ask was a session lead.

Session leads lined up at the front of the room and were given 2 sentences to describe who they were and what they wanted to talk about. They had a facilitator who kept them honest. There was only one female session lead out of about 30 sessions.

I went to a session where the description included the term ‘user centered’ but it was more about identity issues and users being in control of their info rather than about user centered mashups. A couple of efforts in the digital identity space are:

• Cardspace (MS framework to manage digital identities)
• Higgins project (Identity management in Eclipse)

There was a fair level of concern among people at the session about the security of individuals’ data and the content that individuals create (e.g. photographs). Perhaps DRM should be used by individuals on their personal data and the media they create. There was also some belief that if phishing gets out of control, it could kill the web

Silverlight / flash light brought up and discussed as alternates to the browser for mashups. Java wasn't discussed much.

Access is a Japanese company that has ported browsers to mobile phones, smart phones, PDAs, copiers, printers, home apppliances, IP phones, Multifunction phones, video game consoles, musical instruments, TVs, GPSes, set top boxes and business terminals. They partially view the browser as a UI toolkit, and use it to implement applications on different devices. The idea is that if they get the browser running on the device, then its a well understood platform for implementing UIs. They use GCC, linux and sometimes use Java.

Mashups tend to combine data from two or more websites. Some websites have published APIs which the mashup developers use. If a website has useful data (e.g. quote.com for stock prices) but no API, then the mashup developers use screen scraping (navigate the DOM to figure out what data the application needs.)

Kapow’s tool for point and click (and programmatic) web scraping.

There are a lot of location based mashups, and they fall into two categories (with many variations in between): mashups that tell you where you or your group are, to mashups that tell you about the a location, or the location you are.

Some examples:

Location based mashup about specific location: “Bobby’s Bio dventure” mashup that tells you what EPA superfund sites are close to any address.

Location based mashup: “Metrosphere for Andriod” provides info about the location you are at, e.g. restaurants, movies etc...

Group Location Mashup: tells you where the people in our group are based on cell phone location (tower triangulation).

Location based Mashup: Map-and-go: info about a location.

You can find more of the photos from the mashup camp here on Facebook

Thursday

(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):

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.

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

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.

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:

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

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.

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 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.

Sunday

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:

Monday

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...

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...

a 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:

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