I'm curious about something.
Normally in UI's, you'd have a "panel" object which is basically a box. It usually outlines things, giving a rectangular area to work within. Most UI's I've worked with have things based off rectangle areas, such that objects can anchor within the parent area, position themselves relative to it etc.
NGUI in this respect seems entirely weird to me, in that panels by default have no rectangle. What's weirder still is that they CAN have a rectangle, but then that rectangle isn't sizable like the sprite is. Having grab handles for rectangle areas like panels seems to me to be a no brainer. Its there for sprite but not for panel and I can't for the life of me think why?
Having to have an anchor script also seems weird, given most UI's simply have an anchor property that anchors a child in relation to a parent rect (which means of course parents must HAVE a rect).
Can I ask about the design thinking for NGUI? I mean why not follow the conventions of 100's of other UI libraries? Is there some advantage to this method of layout I'm missing? because trying to do things like scrollable lists seems needlessly vague to me this way, or even relatively simple controls seem complex (i.e. I create a button, usually the button would be the rectangle the button covers, rendered with different states, yet with NGUI the "button" is simply a collection that isn't really even loosely connected, so trying to create a button of a certain size means editing colliders and sprites and labels separately).
Perhaps there's an advantage I'm not encountering?
Genuinely curious.