When I started to know GUI in the early 1990s there were always the same type of control to perform an action of the same type, even though the action was different as in
[OK] [Cancel] and
[Yes] [No]. I’ve seen another pattern evolve here which at first doesn’t make any sense at all – but the more you think about it – the more sense it makes.
Here on UX.SE there are the possibility to edit posts, and when you do edit and are done, or decide not to continue the edit you have two options. The first, most visually prominent action you can take is the “Save Edits”-button. The second action are the “cancel”-link with lower case initial letter and much less prominent than the button.
The reasoning behind this could be that the most likely action to take would be to save the edit – thus that action is much more prominent than the cancel action. And to guard further from letting the user unintentionally cancel an edit – there is a follow up Confirm Box asking the user again to abandon/cancel the edit.
But this is speculation, and I have no hard evidence that this is the case. It looks very promising and I’d like to use it in designs to come – but I want to have some proof of the reasoning behind this new pattern. If there are articles, test results or even research available – I’d like to know.