This site is about: (1) my professional self, (2) my research into cognition and (3) musings about the intersection of cognition and design.
Jason H. Wong
Basic cognitive research is a necessary component of successful user-centered design. Only through scientific thinking can we make technology intuitive and productive. My goal is to integrate basic research with useful applications.
Microsoft Packaging Craziness
A very funny video circulated a long time ago asking the question, “What if Microsoft designed the iPod box?” The proposed re-design that followed Microsoft’s packaging guidelines was HILARIOUS and embedded here for your viewing pleasure:
While the Microsoft Vista box is not as ugly as the Microsoft I-pod box, it apparently is hard to open. Hard enough that it warrants a help page on the Microsoft site that includes the 3-step instructions on how to open the box. There are pictures involved, thankfully, since the text instructions are ridiculously confusing:
1. On the top of the box, cut along the grooves on either side of the Microsoft Certificate of Authenticity label.
2. Peel the red tabbed label off the front of the box and discard.
3. Holding the box with the Windows logo facing you, grasp the red tab on the top of the box, and pull it to the right to open the box as shown here.
This is an excellent example of why we need Human Factors.
From the January 2008 issue of Wired
FOUND: Artifacts of the Future
Or, as I would subtitle it: “A Human Factors Nightmare.”
All kidding and craziness aside, heads-up displays are coming into cars very quickly. Right now, they only display speed. The goal is to put information right in the driver’s field of view to minimize eye movements away from the road, similar to the Honda Civic dashboard. This makes sense from a cognitive perspective, except for one major flaw: people can only pay attention to one depth at a time. Therefore, attention must shifted from the road (relatively far away) to the windshield (much closer) in order to glean necessary information. That still leads to performance problems. What some systems are trying to do is to project the information “into the world” so that the speedometer is still on the windshield but appears on the road, so you don’t have to shift attention in depth. Smart.
Of course, while it’s easy to argue that the dashboard is way too cluttered to make driving safe, complex GPS systems are already part of high-end cars. They can be complicated to use and certainly provide a distraction if they make a mistake. Presumably, you’re driving in unfamiliar territory, and your map breaks, forcing you to fiddle with it?
There is a point of balance, where more information helps the driver drive better or multitask. But this makes human factors research all the more critical to ensure safety at all costs while not making driving an unbearable task.
More is not always better!
Microsoft Office has been accused of bloat for many years now, and Microsoft Word has been no exception to this criticism. To some extent, this is unavoidable: Word needs to be such a general-purpose product that you have to cram it full of features in order to get as many people to use it as possible. Of course, each of those users only needs a subset of those features. I use the Word Count feature when I’m writing a manuscript for a journal that has a strict word count, but most users have no need for it. For creative writers, a majority of the features in Word are useless - tables, graphs, even fonts - are unnecessary and clutter up the creative process.
Steven Poole has written an excellent blog post about his history with electronic writing and promotes a program called WriteRoom that does away with all the distractions and interruptions that Word and computers in general provide. Writing suddenly looks like paper again with WriteRoom:
This application would likely drive me crazy, as I often need PDFs and spreadsheets open as I write my scientific manuscripts. But for those that just need to write, the distraction-less interface of WriteRoom is likely as close as you can get to computerized perfection.
Incredibly long menus in Excel 2007
I do a lot of data analysis in Microsoft Excel, and I’ve been using Office 2007 for some of it. I like my spreadsheets to look attractive (and be organized), so I use the Borders function a lot to delineate different pieces of data. In older versions of Excel, there were a limited number of buttons to draw borders around groups of cells. In Excel 2007, when you click on the Borders button, you get:
Not much to say except for: Wow. Bottom borders, top borders, top AND thick bottom borders. Oh, the choices! And just in case there aren’t enough choices, there’s a More Borders… button at the bottom. Thank goodness. OK, all snark aside, let’s analyze this design.
Why did the designers make a menu this long?
You need to give your users options - options are good in software design. However, in this case, every possible option is a staggering number of possible permutations. You can’t include every possible one, so you should make the most common options more accessible (by being in the menu) to speed up the selection process.
What went wrong?
Well, you can see what went wrong. So many options! Bottom Double Border, Top and Bottom Border, Top and Thick Bottom Border, Top and Double Bottom Border… Draw Border, Draw Border Grid…. the list goes on and on and on, and it doesn’t seem to have any sensible order. Clutter makes visual search more difficult, and working memory is unable to remember more than a few items. This menu is too cluttered and too seemingly arbitrary to be good design. What research was done at Microsoft to determine that “Top and Double Bottom Border” was a border configuration so popular that it warranted a spot in the menu?
What should be done?
Make a portion of the menu a box where you can click to selectively add the borders you want. By including a clickable grid in the menu, you present the user with a huge number of options, but it’s not daunting because the user can “draw” the lines that they want. It also wouldn’t be a bad idea to move the “More Borders…” option closer to the top of the menu. If users are not satisfied with the menu options given to them, it’s likely that they never will be. Giving them one-click access to a more complex menu box will satisfy them, and users who don’t need the “More Borders…” box will only have to skip over one entry, which is not a huge cost.
In short: Make the menu shorter to reduce the load on working memory. Reduce clutter to ease visual search. Change the confusing labels of “Borders” and “Draw Borders” - what’s the difference? I don’t know. But making these changes would improve the menu quite a bit. One menu down, only hundreds more to go!
GUI Wars: Window switching: OS X vs. Vista
Most computer users have a lot of windows open at once - web browser, e-mail, Word, Excel, etc. etc. Switching between them could always be accomplished through a variety of methods, but the latest operating systems have tried to jazz that up. Since a year or two ago, Apple’s OS X has had a feature called Expose. You press a mouse button or function key, and all the windows shrink to give you a live preview of each one. Then, you click on the one you want:
This works nicely because it gives you a bird’s eye view of all your windows. You can navigate this screen with your mouse or with arrow keys, so you can stay mouse- or keyboard-consistent.
Doing this in Windows Vista, the newest version of Microsoft’s dominant operating system, looks like this:
Pressing Alt-Tab brings this up, and multiple presses of Tab cycle through all the windows. This is nice because you can flip through your stack of windows until you find the one you want. In fact, the feature is called Flip3D.
So the question is: which one is better? To answer that, we need to delve into the visual search literature. Visual search is a task used over and over again in attention research. You’re looking for a target amongst many distractors. What’s the best strategy you can employ to find it? Well, it depends on the properties of the targets and distractors. If your target is highly salient (i.e., unique compared to all the other items), search can be fast when you’re presented with all of the items at once. This is known as parallel search, and it’s fast no matter how many items are on the screen (find the green square):
However, if your target isn’t salient, then you’re slower. You generally have to examine each item individually until you find your target. This is known as serial search (find the green square again):
So back to the window switching interfaces in both operating systems. How often is the window you want salient? Not that often - it’s not like the window that you want is bright green and everything else is gray. But you do have expectations sometimes. “I’m looking for iTunes, so I expect the window to look a certain way.” So you may be able to perform parallel search and get your search done faster. Not always, but sometimes.
Think about how both interfaces work - which interface allows you to perform parallel search? Expose! Because Flip 3D does not present you with all targets at once, there is no way you can perform quicker parallel search. You must perform serial search each and every time you switch windows. Considering how much people multitask, this can add up to a meaningful amount of time quickly.
There is the issue of clutter, though. More clutter in a display means longer searches. Expose definitely leads to more clutter than Flip3D, so a search in Expose may take longer. This primarily occurs is you have many windows open. Flip3D is not as cluttered because it only presents you with one window at a time, which underestimates how many separate objects we can handle at a time.
All in all? When it comes to switching windows, OS X’s Expose has the science behind it. By allowing users to engage in parallel search for their window, multitasking can be done much more efficiently than Vista’s Flip3D.







