Is there any justification for different behavior of different ways of triggering the same action?

Question

Something that drives me crazy in certain user interfaces is when different methods of triggering the same action have different behavior.

Examples:

• In Microsoft Word, one can introduce a bullet point list by pressing the button on the ribbon or with a keyboard shortcut (SHIFT + CTRL + L), but the behavior of the list is different in each case. The list started by pressing a button is indented and uses different bullets for different indentation levels. The list started with the keyboard shortcut has no indentation at the first level, and the different levels all use the same bullet.
• In Chrome, opening a new tab can be accomplished with the context menu or CTRL + click. However, in certain cases, CTRL + click does not actually open a new tab, while the menu does. For example, see the picture bar links at the top of this blog.

Is there any possible UX justification for the same action having different behavior based on how it is triggered?

Given Microsoft's track record, I'm willing to write off the Word behavior as illogical (and probably supported for legacy reasons). However, Google has a very strong track record of user interfaces that are driven by principles and well-tested. I can't understand how the above behavior would be desirable to anyone.

It seems to me that actions should always behave the same, no matter how the user triggers them. Does anyone have any justifications for such differences?

Update: I think the specific case in Chrome occurs when a page blocks new tabs with Javascript. But still, that doesn't justify silently behaving differently in my mind. There are UI paradigms for overriding default behavior, such as asking the user or having an option that controls the behavior.

Answer 1

Is it really the same action?

Some actions will have the same result in many cases, even most, but may differ in some. Some actions will be differentiated by intermediate steps that are optional or impossible one way but mandatory in the other. Such similar actions should be kept separate, though.

Example

If I hit Cmd+N (Mac) or Ctrl+N (Win) in an editor, I expect to create a new document or maybe a new project, depending on the exact nature of the application. I expect the same if I click the blank page or starred page icon, which is usually located leftmost in a tool bar. By convention, the respective command must also be found in the main menu bar.

In many an editor, new documents can be based on any of a number of templates, but there’s usually a vanilla one with minimal formatting or another most-used default template. As a user I’ll want to be able to access either, regardless of my preferred input device, but the default behavior for creating a new document may depend on the way I’m using to access it.

When using a keyboard shortcut one may assume that I intend to keep my fingers on the keys and continue working as soon as possible; I’ll also know other shortcuts to do stuff quickly – I’m a power user. When using the mouse to point and click on an icon, however, I’m already prepared to select an option from a long or graphic list – I’m mentally and physically in Dialog Interaction mode, not Command mode.

Therefore it’s reasonable to make ‘New document from template …’ available with a more complex keyboard shortcut like Cmd+Alt+N or Ctrl+Shift+N and also make ‘New document’ available as a hold and drop-down option (or similar) for the page icon. Both will be listed next to each other in the fallback menu bar.

In conclusion, the action for the default, simpler key press does not exactly correspond to the action for the default, simpler mouse click, but the end result is similar and there exists an alternate, more complex way to do the respective other action with every means of input. The reason is that the input device to invoke a certain family of actions tells something (but not everything) about the user’s proficiency and state of mind.

Answer 2

In the two examples you mention, I'd say that the short answer is: no it's not justified. There's no clear purpose of the different behaviors, and there is nothing to hint to the user beforehand that the behavior will be different. When it comes to interaction we really don't want to surprise the users (except some cases where we want to pleasantly surprise them, but this should be a surprise regarding content or visual design not interaction).

Reference: See for instance Consistency and standards in Nielsens 10 heuristics for User Interface Design http://www.nngroup.com/articles/ten-usability-heuristics/

Are there other examples where we actually want different behaviors for different ways of triggering actions? Absolutely! A common example is if I've triggered an action with a special shortcut or from a particular view. Clicking a delete-button should perhaps normally give me a confirmation-popup, but if I'm in super-admin-mega-delete-mode perhaps the same action should not trigger a confirmation due to the purpose of that mode (efficiency). This would fall into the category Flexibility and efficiency of use from the same list of heuristics.

It is, however, important to note that even if we're adding actions and shortcuts to increase the efficiency for our expert users, it should still be clear which behavior we can expect from these shortcuts!

Answer 3

Two possible reasons come to mind loosely related to examples above

Default vs. Informed Selection

With the mouse driven UI user can select from a presentation of bullet styles. This selection allows for informed choice. IIRC Word remembers last choice made by mouse, but the full set of choices is adjacent and default is highlighted.

With a keyboard input the expectation may have be that this always maps to one base default bullet style. Rather than producing a default style that you can't visually identify or control.

Heuristic at play is Recognition rather than recall

Input Device

Some operations are harder to achieve with one input device over another - thus the flow may adjust to provide best path for a given input device. e.g. Switching tabs is easy with a mouse, so leave it as user action, but harder with keyboard, so take user into new context immediately.

Answer 4

Is there any possible UX justification for the same action having different behavior based on how it is triggered?

Yes! The UX justification is context-awareness and adequacy of tasks.

The use of mouse or keyboard is different in your actions you can perform (i.e. preciseness and speed). If you interact by speech interface you may describe your desired action or you may need a "action detour". Same with using a website on mobile or desktop, it may need different grades of details in your actions (or prefilled templates).

One may argue, that user profiles (UX justification individualisation) like Googles search outcome is a different behaviour, too. And it's okay for the user. Even here it depends on device (aka cookies) and interface (like Voice Search).

So, the same action may have different behaviour based on how (interface) and with what (device) it is triggered.