Oh there's more than just the original three? Interesting. But my "don't click the links" wasn't specifically for you, but for any forum reader.
You've read them?
Ah.
Hasn't everyone? For a while they had almost Tolkien-like status among the reading public. But let's not hijack this thread...
Huh. I've only recently read it, from the library.
I wonder what that would look like... (the "That could even be written in Snap*!*, since we have fast hyperblocks." part)
All three books were #1 bestsellers as they were released. The entire reading adult population of the world seems to have read them! (I don't think parents let their kids read them, for obvious reasons.)
Oh, you know, supposing the thumbnails are 10x10 including space in between, which isn't realistic but makes the arithmetic easy, to get row 2 column 3 you'd say
roughly speaking, and of course those four numbers would be expressions, not constant.
So, for pixel (3,2) in the thumbnail, would you average all the values of that?
No, no, the result of that expression would be a list of the 100 pixels making up that thumbnail.
But now that I think of it, I'm confusing two things, the thumbnail png (of which there's only one) and the actual costumes (in which the full set of pixels is stored). So there's really no need to extract one thumbnail from the png. We'd just look at exactly where the user clicked the mouse and load the corresponding costume. But I guess to make it look exactly like the existing interface we'd have to draw rounded rectangles around the selected costumes in the thunbnail png and not import them until the user clicks the button. Drawing and undrawing those rounded rectangles is where hyperblocks would probably come in.
I don't understand. Why would you have only a small area of the costume for a thumbnail and not a downsized version of the costume?
The costume would be how it's always been.
The costume chooser, which now has to read every costume and generate a thumbnail for it on the fly, would instead just load one picture, the array of thumbnails, pre-generated programmatically and just saved as a big .png. You'd click on it, the code would translate mouse coordinates to which costume you want, and load it.
Oh! I thought you were talking about generating thumbnails for costumes. But now I get what you're talking about.