In Comparison: Multiple Selection, part 3
Parts 1 and 2 of this series of posts were the result of my exploration to determine the behavior I would specify for a new, browser-based UI. Now that I’ve explained how it works in Windows XP and Mac OS X Lion, it’s only proper that I wrap up by relating my own design rational.
I sat down with the developer who is to implement the feature and went over my findings. She agreed that including an item that was just deselected in a new ranged selection seemed like a bug. While I like the flexibility of Apple’s implementation, I agreed with her that for our purposes, it was needlessly complex. Together, we settled on a single anchor. We decided to set the anchor on clicks and shift-clicks and to not move the anchor for control-clicks. We also decided that, to keep things simple, we would never deselect items on a shift-click. This goes against the standard desktop behavior, but my thinking is that the average user doesn’t likely realize the intricacies of such interactions. As long as we do something that is sensible and internally consistent, I doubt it will be noticed.
I documented the behavior in a series of Interesting Moments Grids, like the one below, that the developer will reference during implementation, as well as during testing.






The next character I type gets the acute (é). There are corresponding shortcuts for all of the standard accents, typically assigned to the letter that most often receives them. The tilde, for instance, is option+n (ñ). The umlaut is option+u (ü). This makes them extremely easy to remember—much easier than seemingly random, four-digit codes—and it is the same combination regardless of case. To get a capital, you just hold down the shift key when you type the letter, just as you normally would.