I tend to give the browser a certain degree of latitude in Web design before and after the optimal resolution. The big move is to make full use of the screen space of the large screen users in order to improve the experience, taking into account that sometimes the user does not maximize the browser window or fill the entire screen.
For example, when I work, I often have two Windows side-by-side, one window is Google Docs, and the other window renders references. If you're working in a café, it's just a laptop. A screen, each window 700px wide, many times I have to constantly drag the horizontal scroll bar, very tragic.
What if the browser's window is particularly large or very small? I think that can not be optimized for these extreme situations, but the bottom line is not to make people feel wrong, but also to the Web site to reserve a fallback to the interface. Generally, I give the interface a maximum width and a minimum width, so that the width doesn't get out of hand. Like IGoogle on a small screen, it should not:
These are certainly not the mainstream users face the scene, but designers should not be the pursuit of perfection. Here is another factor is that the adaptive width of the page design and development workload is greater than the fixed width, many people may have to consider the value of this additional input is not worth the problem, such as I will be the width of their blog to write dead, anyway, no one to see ...
Most of Google's product interface is adaptive browser width, the product is different, the specific approach will be different. For application products, the interface framework is generally adaptive, but in order to ensure the readability of the text, the text content area will often be limited to a maximum width, like Google Reader:
New Google News Homepage The entire fixed width center, in addition to the above unified navigation bar (I think can do more flexible point):