https://bugs.kde.org/show_bug.cgi?id=488828
--- Comment #3 from fanzhuyi...@gmail.com --- There seem to be two issues raised here: 1. window sizes being too small, and 2. windows moving from their original positions. I will address each of these in turn below. 1. window sizes are too small The new algorithm should be pretty efficient in using the screen space. In fact, it should be more efficient than the old algorithms than both the old algorithms in using the screen space. Hence, if window are too small, it's because the margins are too big, or because there are too many windows open (we can't do anything about the later). For the former, it would be straight forward to decrease the size of the margins if that indeed is the issue. Could you please upload a screenshot showing the problem you preserve, so we could take a look? Also, letting us know the resolution of the screen would be helpful. 2. windows moving from their original positions. IMO if we operate with a tractable (i.e., computationally efficient) algorithm, there is a fundamental trade-off between making efficient/aesthetic use of screen space, and minimizing movement. (if there is an efficient algorithm that can accomplish both, I would love to learn about it) The old natural algorithm often made horrible use of screen space, and people complained about that as well (e.g., BUG 477833, BUG 477833, and attachments in BUG 450263). Hence, at some level, we have to make a judgement call between minimizing movement, and making efficient/aesthetic use of screen space. Users should have little need for this algorithm when all their windows already don't overlap. Thus, we decided to some what prioritize making efficient/aesthetic use of screen space, given that the algorithm is most useful when a lot of windows overlap, and in this case minimizing window movement doesn't really matter. In addition, the implemented algorithm has components to minimize movement, contrary to what I said in that comment you mentioned. In the merged version, (assuming a landscape display), the windows in each row are sorted by their original horizontal positions. Hence, horizontal movement is minimized. If you could upload a screencast demonstrating the problem you see I would be happy to take a look. -- You are receiving this mail because: You are watching all bug changes.
