I've just experienced the wholly pointless exercise of emerging bootsplash-themes on a netbook with a 1024x600 screen. 176 downloads, 64MB in total, none of it usable. I'm not bitter, but it strikes me that this could perhaps be avoided.
What springs to mind would be a class of USE flags similar to INPUT_DEVICES, ALSA_CARDS etc called, say, FBCON_RESOLUTION. The user would then specify only the supported resolution(s?), and the ebuilds for bootsplash-themes, splash-themes-gentoo and splash-themes-livecd could respond to this by only installing themes that support that resolution.
Of course this alone would mean a zero-file install in my use-case for bootsplash-themes in particular, so a warning or bailout would maybe be the answer in such scenarios. I'd welcome insight on the correct approach here.
A second possible approach would be to reorganise the ebuilds such that there was one for each resolution (somewhat like some font packages), which would mean less deadweight for the user, but appreciably more effort to maintain so many more ebuilds.
What do people think of this? I'm prepared to undertake the work, but would appreciate opinions on (a) whether it would be an acceptable strategy in general, and (b) what the preferable method of implementation would be.