But those don't easily scale to different dpi and layout sizes. There are non-standard ways but nothing has taken over. So flat icons and solid colors it is.
For a long time, designers/artists might have produced several versions of the "same" icon at different sizes for use in different contexts. Sometimes you'd have a large area to work with, maybe 128x128 or 256x256 pixels. You could have a lot of detail at those sizes, in whatever visual style you liked. At smaller sizes, you didn't just scale down the large icon mechanically. Instead, you'd redraw it to preserve the essence but remove finer details and simplify the overall design as you had fewer pixels to work with.
A similar idea existed in type design. A comprehensive font family might include different "optical sizes". A caption font designed for use at small sizes might have lower stroke contrast, while a display font intended for use with large headlines might feature finely detailed serifs that wouldn't work well at small sizes. If you took the caption font at 8pt and the display font at 48pt and scaled them mechanically to be the same size, they'd look quite different.
In both cases, the advent of mechanically scalable image formats like SVGs and most digital font formats marked a change in common practice. Now you only needed one file to work at any size you wanted, but the subtlety I was talking about above is lost if you're no longer creatively adapting the same basic design for different contexts but instead you have to draw a kind of lowest common denominator that works tolerably at all sizes.
As others have observed, this shift certainly makes things cheaper, but potentially at the expense of quality of design. We only need to look at the bland, homogenous nature of sites built with flat design and its derivatives, not to mention all the obvious usability problems, to see that the popular approach right now isn't actually very good.
For a long time, designers/artists might have produced several versions of the "same" icon at different sizes for use in different contexts. Sometimes you'd have a large area to work with, maybe 128x128 or 256x256 pixels. You could have a lot of detail at those sizes, in whatever visual style you liked. At smaller sizes, you didn't just scale down the large icon mechanically. Instead, you'd redraw it to preserve the essence but remove finer details and simplify the overall design as you had fewer pixels to work with.
A similar idea existed in type design. A comprehensive font family might include different "optical sizes". A caption font designed for use at small sizes might have lower stroke contrast, while a display font intended for use with large headlines might feature finely detailed serifs that wouldn't work well at small sizes. If you took the caption font at 8pt and the display font at 48pt and scaled them mechanically to be the same size, they'd look quite different.
In both cases, the advent of mechanically scalable image formats like SVGs and most digital font formats marked a change in common practice. Now you only needed one file to work at any size you wanted, but the subtlety I was talking about above is lost if you're no longer creatively adapting the same basic design for different contexts but instead you have to draw a kind of lowest common denominator that works tolerably at all sizes.
As others have observed, this shift certainly makes things cheaper, but potentially at the expense of quality of design. We only need to look at the bland, homogenous nature of sites built with flat design and its derivatives, not to mention all the obvious usability problems, to see that the popular approach right now isn't actually very good.