It is possible, for instance, to get a view of the directory structure outside of your home directory, but the means are buried several clicks deeper than in the GNOME 2 series. That is fine for regular users, but less than ideal if you happen to double as root user and want to refer to the directory hierarchy while you have a terminal open.
Just as important, if you open too many windows at the same time, the bottom parts of the menu become inaccessible. Since Recent Items are at the bottom of the menu, the only way to take advantage of this feature is to close applications and reduce the amount of space occupied by the favorites list -- an interruption that distracts from the simple act of trying to start an application.
Other problems with the overview menu include the list of applications, which is alphabetical, but not grouped by categories. If you want to find, for example, administration tools, you have to scroll through the entire list. Similarly, because favorites and open applications are in a single display, you cannot always be sure which is which.
Presumably, such quirks will be smoothed out before GNOME 3's release. Meanwhile, familiarity does mitigate them. Yet, as you deal with such limitations, you may feel -- yet again -- that you are working the way that GNOME Shell requires, rather than how you would prefer.
The preview, I suspect, shows GNOME Shell in its purest form, with the division of labor between workspaces and the overview rigidly observed. By the time of GNOME 3.0's official release, perhaps applets and more customization will soften the distinction between the two views, and make GNOME Shell more friendly. Perhaps, too, the overview menu will become more flexible.
Meanwhile, I hesitate to pass judgment on a work still in progress. GNOME Shell offers plenty to admire, particularly its ingenuity and its treatment of workspaces as an integrated part of the desktop, rather than as an add-on (which is how most desktops treat them).
Yet, without more flexibility, GNOME Shell requires considerable planning and organization to make it a tolerable environment. Once you have your standard workspaces set up according to tasks, then everyday computing becomes far easier and more tolerable. But, even then, GNOME Shell requires far more customization than the GNOME 2 series. You have to wonder whether uses will accept this requirement, or switch to alternatives that are more ready to use out of the box.
My growing conviction is that GNOME is on to something promising with GNOME Shell, but hasn't worked out all the implications of the design to make it acceptable to different styles of computing. How many users greet GNOME 3.0 as the future of the desktop and how many immediately demand a fork of GNOME 3.0 will depend very much on how well the GNOME project addresses the need for greater flexibility. But, because GNOME Shell is so different from the GNOME 2 series, I suspect that we will most likely see both responses.
ALSO SEE: Eight Ways GNOME Could be Improved