Those of use who routinely use SSH know how sheerly infuriating it is when the agent asks you if you want to trust a host key and will you please enter “yes” or “no” (not “y” or “n”, mind you) to confirm your choice. Yet, none of us would like to change that because, in the end, it is for our good.
SSH knows it’s annoying. It know it forces us to go out of our ways, think about a word and type it on our keyboard to interact with the program. But, by doing so, it also knows that it forces us to have a meaningful interaction with our machine: even the less experienced of users will realize, by doing so, that he provides an answer to a specific question.
The question SSH asks is relatively cryptic (Do you want to trust a specific host key?) but dismissing it without thought is impossible. Modern dialogs, on any interface, come with a default “OK” button that one can systematically trigger by pressing return or enter on our keyboards. It’s easy, tempting and a seemingly quick solution.
But maybe it is too simple? After all, in many cases, security revolves around a couple dialogs that all sport the same shiny, pulsating/underlined/colored “OK” button that we all want to click on to just regain control of our machines. Look at the extents to which Apple has gone when you enable FileVault on a Mac: there are two dialogs, both magnificently breaking the Aqua guidelines by piling up colored text, bold lines, buttons and warnings in just about every direction. Why? Because these are required to shock users and get them to think.
Maybe a simply text field, along with the instruction “Please type Yes or No in the box below to confirm or cancel your choice” would do wonders. It certainly would be “new” enough for users to think and feel like they’re about to make a big decision but it would keep dialogs clean and to the point.
I’ve always wondered how the community felt on this topic. Am I the only one thinking text dialogs have their advantages?