I would be interested to know if anyone has explored reduction in potential
visual overload of graphical based interfaces? Particularly around action
buttons which take action on some visible object in the UI.
There are two aspects which I'm considering, one is the visual appearance of
the button (should it look like a button - affordances - or can it appear
much simpler). The second is whether it can be hidden altogether until the
user has moved focus onto the object which the command applies. This would
mean the UI would show all the objects, represented by their important
content whether that be a graph or a row of numbers or a persons
photograph. Then when the user focuses on an object, the UI presents the
actions you can take on that object. For example when you hover over a
contact in gmail it presents the option to chat or phone that contact. Or
in http://www.tiddlywiki.com/ where when you hover over a posting, relevant
actions are presented along side the posting such as view the posting. This
makes it easier I think to read the content without presenting the same
actions for each posting.
I guess my concern was with discoverability of what you can do with the UI
and issues with consistency. For example (maybe a bit extreme) if you were
in an office with doors to meeting rooms and until you approach a door its
handle is hidden, when you get close enough it appears.
I'm currently looking at a table which offeres the ability to edit, delete
and copy elements of the table and was considering only showing these
actions when the mouse hovers over an element.