MobileFrontend is, as much as some of us like complaining about it, fundamentally a good idea (actually that's part of why we like complaining about it - we'd just be ignoring a bad idea). We should chop it up and reorganise it to optimise its effectiveness:
- Move the mobile detection and setting skinclass for mobile into core
- Turn Minerva into a standalone skin that can be used for desktop, mobile, or both
- Redefine MobileFrontend as a general extension for simplifying and optimising potentially any functionable skin for mobile (hiding things that simply do not apply to mobile, rejigging the table of contents, thumbs, section handling, infoboxes and templates, special pages, etc)
There are probably other features I completely missed that could go wherever, but you get the idea. I would like to see this. It seems to already be going in this direction in some ways. Will it actually happen? Can we agree? Why is it skinclass instead of just a skin name? Where is it going?