I think there are two separate issues with the controversial move of the window control buttons to the left.�
On the whole I am not averse to it, and I believe the logic described here�
http://www.ivankamajic.com/?p=281�(namely that most menus are on the top left etc.) make sense.
However, the decision to place the close button on the inside (i.e. third of the three buttons) strikes me as very very odd. While the logic of not placing the most destructive action first makes some sense, it also makes sense to me that the most significant action that can be taken with respect to a window (close it) should be in an extremity. So I would have placed the close button in either the left or right corner, even if this meant that the most destructive action would be first. Although keeping the minimise and maximise buttons on the left and moving close to the top right corner overcomes the latter, it has the disadvantage of splitting the three controls and increasing complexity.
I wonder if the design team could share more of the thinking behind this and also let us know whether user testing was conducted in order to weigh up the various options based on some real usage data.
My analysis of the feedback has it falling into more than two categories so, let's compare:
1) The theme that is foremost in my mind is the fact that Lucid is an LTS. �
We deployed in Alpha 3; we are monitoring the feedback and will make a descion about whether to keep the current order or tweak it in the next few weeks.
2) Learning curve
This is less of an issue than I had anticipated; after a very short amount of usage, neither button position nor order have any significant impact.�
3) Compatibility with Mac OS and Windows
We are designing solutions for Ubuntu. Windows and Mac OS can make changes. My father was always quite intent on reminding me that he was only ever interested in what I thought or did, not where I stood in relation to my classmates. In running usability tests with first time users, things like button order are quickly shaken off - no access to their iTunes library is a far bigger stumbling block in people's minds!�
4) Resistance to change
This is completely unavoidable and to be expected.
5) Button order
Is based on least destructive first (when reading left to right: cultural differences noted.)
6) Fitts law, etc
In devising tests the most important thing is to decide what you are measuring and why that matters. �For example, citing Fitts law (which I realise you don't mention but others have) is useful if one argues that time to button is the factor which is of most value to the user. If time matters most then the mouse is rarely the answer (
http://en.wikipedia.org/wiki/GOMS).
With design, you can't always rely on standards and measurement and need to push for innovation. The place where opinion is unavoidable is where one should draw the line.�
Kind regards,