News for hackers. Stuff that matters.
Re:Unanswered user makes question their MS thesis and answer... by TazeTSchnitzel on Sunday January 03, 2021 @02:59PM (#25624668)
Do popular GUI frameworks that present an immediate-mode interface actually use an immediate-mode implementation under the hood?
Re:Unanswered user makes question their MS thesis and answer... by mst on Sunday January 03, 2021 @06:01PM (#25626101)
"Dear Imgui" is pure immediate mode both interface and implementation wise so far as I can tell.
Re:Unanswered user makes question their MS thesis and answer... by dkersten on Sunday January 03, 2021 @06:37PM (#25626358)
Although it is worth noting that calls to Dear ImGUI don’t directly render, but rather build up state to batch render later with your graphics API of choice.
Re:Unanswered user makes question their MS thesis and answer... by mst on Monday January 04, 2021 @01:31AM (#25628408)
Fair.

"blitmediate"

Re:Unanswered user makes question their MS thesis and answer... by ma2rten on Monday January 04, 2021 @03:22AM (#25628849)
The Windows API asks the application to redraw the window on certain events (e.g. resize).
Re:Unanswered user makes question their MS thesis and answer... by anderspitman on Sunday January 03, 2021 @03:28PM (#25624927)
The most popular immediate mode system I'm aware of is React, which definitely uses retained under the hood.
Re:Unanswered user makes question their MS thesis and answer... by voodootrucker on Sunday January 03, 2021 @06:41PM (#25626388)
As a person who has used a slew of UI frameworks and written one or two: All of this is happening at a layer considerably lower than React.

This is what's happening when the browser redraws when you scroll for example. Arguably it has to do with the DOM, but certainly not how the DOM gets updated.

Re:Unanswered user makes question their MS thesis and answer... by mst on Sunday January 03, 2021 @06:02PM (#25626115)
I think I'd argue that React is immediate mode with caching and then a retained-ish driver layer to talk to the underlying DOM.

But given the constraints of the "talking to the underlying DOM" part it's going to be a bit handwavey no matter how you argue it, I suspect.