It came up in irc that having a fixed master concept is a bit of a hack. We really need more general mechanisms for tiling in the default 'tall' (or its mirror) mode. The default mode is too ad hoc ! :-) So, um, there's probably been UI research on this, right? Anyway, some interesting papers to read, http://www.cs.umd.edu/hcil/elastic-windows/ https://drum.umd.edu/dspace/bitstream/1903/439/2/CS-TR-3522.pdf http://citeseer.ist.psu.edu/kandogan96elastic.html Also, "Vanishing Windowsa technique for adaptive window management" And from there we can probably find some other background reading for new tiling ideas. Removing limitations on what tiling forms can be achieved will be one of the main directions for the next release, I hope. -- Don