TechEd Day 2: Interface Design Patterns

DSCN1539
This morning at TechEd Australia 2007, I attended the “Interface Design Patterns” session by Darryl Chantry, an architect on the Architecture Strategy Team at Microsoft (Australia). A session mostly directed to developed on how they should start thinking about designing application for the user, instead of themselves.

DSCN1546
The many and overlapping disciplines of design. Darryl places operating systems under human-computer interaction.

DSCN1547
The many ‘approaches’ to tackling design problems. User-centered design is becoming popular, but may not be the most appropriate because users might not know what they ‘want’. Darryl uses the Google homepage as an example of activity-centered design, where the search is the activity and simple search box on the homepage. He adds most developers use the genius design approach, whereas developers define what is right for the users.

DSCN1548
Darryl told a story about a scenario at a Microsoft research activity in Redmond. The test involved two groups of users both using a poor quality PC for a week. At the end of the week, one group was asked to answer a survey about their experience on the same crap computer they’ve been using, and the other group answering the survey on another and presumably better computer.

They found the results were at two opposite ends of the scale, where the group who used the same computer for the week and survey were more forgiving about the experience and other group more critical. From the later interview they found out the forgiving group had ‘feelings’ for the computer and didn’t want to “hurt it’s feelings”.

DSCN1549
Sometimes, designers have to be knowledgeable in many many disciplines. For example in the above technical drawing explaining a medical condition, the designer would not only have skills in drawing human figures, but knew of both the human skeletal system and the nerve system.

DSCN1550
Then we watched the TEDTalk video about BumpTop. Looks like Darryl is a TED fan too.

Microsoft pie menu patentMicrosoft pie menu patent
A specific feature he pointed out from the BumpTop video was the use of
circular or pie menus. This is a user-interface concept Microsoft has patented over and over again, so I’m not surprised if they were to go ahead with the radical idea. The biggest benefit between a pie menu and a traditional drop-down menu is the ability to select any of the available options with the same amount of ease. Darryl called this “ergonomic value of design”, where the interface actually reduces RSI.

DSCN1559
Moore’s Law states computing power doubles every 12 months, but we haven’t been able to touch any faster. User interfaces haven’t changed to increase user efficiency as much as computer efficiency. The question Darryl asked was,

Which takes longer?

  1. Heating water in a microwave for one minute and ten seconds.
  2. Heating water in a microwave for one minute and eleven seconds.

DSCN1560
Of course, this was a trick question. The answer was “one minute and eleven”. This is because of the way microwaves have been designed with the number pad buttons. From research, they found pressing 1-1-0-OK takes three seconds longer than pressing 1-1-1-OK.

DSCN1561
Redesign gone wrong can costs business millions of lost revenue through user dissatisfaction. Better to realign, strategy and priority. This is what we should expect more of from Microsoft in the future. Instead of “big bang” redesigns, we should see smaller tests in the market with new ideas.

DSCN1564
Continuous improvements than one big change.

DSCN1566DSCN1567
Technologists sees many layers within systems and applications, whereas end users usually just see one.

DSCN1569
“Think systems, not products.” The story about how the iPod owned the Sansa.

DSCN1570
The Sansa was technologically superior, it had better sound quality, better connectivity, more memory and costed less. However, less people own a Sansa. Because music acquisition is a very hard thing, Apple got it right. Apple built an entire system, putting the right technology in the right place.

DSCN1574
The Molecular Experiment did a case study with a fake wine store for usability studies. Some interesting they found including users ‘want’ to see a loading screen. Waiting for a few seconds, users feel like they are in control – “this interface changed because I did something”.

For some reason, users also expected the back button in the browser to actually undo form actions on the website. To work around the problem, Darryl claims Molecular actually built a framework which overrides the browser’s back button to do custom functions. They will also be building in similar functionality in Silverlight.

DSCN1578
Additional resources:

DSCN1579
And, “more additional resources”, in case additional wasn’t enough.

7 insightful thoughts

  1. I very much enjoyed this article, but I have to ask; why is it the only thing Microsoft has/is willing to talk about is a circular pie menu? Perhaps there was more and I am merely picking out one topic from a basket of 100, but nonetheless I can only wonder what is going on in Redmond R&D.

  2. This is a gold commentary of an excellent session. Wish I was at Tech.Ed šŸ™‚ It’s good to have you on the ground to cover the event for the masses.

  3. “designing applications with the user in mind”… Wow… has Microsoft only found about that now? Perhaps that’s the reason why Google, Apple etc. are running circles around them.

    makes you wonder how they spend all those billions of research dollars.

    At least one Microsoft person at last aknoledges that the iPod/ iTunes experience is easier to use than anything offerd by competitors.

  4. My microwave has buttons for 10 minutes, 1 minute, 10 seconds and 1 second. So to get 1:10 I press 2 buttons and for 1:11 I press 3…

  5. Fascinating stuff. šŸ™‚

    The circular pie menus are hardly something new, though they may be new to Microsoft; I’ve seen them used in a number of games over the years, not the least of which being Linden Labs’ Second Life.

    The problem with pie menus is that real-estate is limited, so using them in a case where a large number of options is available is rather tricky. Also, I’ve never seen an implementation that handled submenus well; pie menus just don’t cascade into submenus as naturally as a drop-down menu. Just try to imagine the typical context menu you get when right-clicking a file in Explorer; it’d get ugly pretty quickly.

    About “Loading” screens, what users really want is to know that ‘something is happening’ when they click on something; they know that the computer is /actually/ responding to their input. I frequently see users start up multiple instances of a single app because said app has no ‘loading’ or spash screen; they don’t see anything happening, so they try it again and again, then get frustrated when their computer slows down and multiple instances pop up.

  6. Kerry is spot on with both the comments on pie menus and Loading screens.

    Pie menus, or radial menus, with submenus are a nightmare to navigate. They are like those web pages which have tiny menus that slide out, but only give you a tiny amount of room to maneuver and if you go outside their tiny boxes they reset. Designers really need to use software the way real people do it and stop trying to be all design and no function.

    Loading screens and their equivalent are the bane of my life. A number of people at work will not wait for an application, eg Excel, to finish what it is doing. If they have to wait for more than 10 seconds, they will open up Task Manager, see that the status says “Not Responding” then kill that application. They don’t realise that it can just mean the application is busy.

Comments are closed.