I agree with TesTeq that going through the whole next actions list(s) multiple times a day, as per vanilla GTD, is not ideal. And, let's be honest, vanilla GTD is not crystal clear (not to me , anyway) on how to best deal with that, and this, in turn, is one of reasons why so mnay of us have our little personal tricks (be it time blocking or highlighting or flagging or scheduling ...).
I also agree with GTD-Sweden that if you are prepared to scratch and redo your "today" list anytime during the day, you are in fact flexible, and you have easy access to a handy menu of highly "likely choices" for the day, and if you need to change it it is not too much work to just "unstar" ("unflag") these few actions (or add a few more). (I use that method myself, and in addition I can always dive into a particular context, project or area if I want to focus on that all of a sudden.)
And I agree with Longstreet that if you put a few really important ones on the calendar (instead of, say, flagging them on your list) you still have a very manageble situation. It all works.
Somehow I feel that the apparent differences between different people's opinions perhaps lie more in the fundamental views we take than in the exact "tricks" we use. At a fundamental level, some of us (e.g. I) feel that whatever we plan will not happen like we planned it anyway (due to interruptions, new events, new ideas ...) and we hate to redo too much of our lists or see the lists all twisted or made useless just because some particular date or time point has passed. I want some form of "planning" that stays as valid as possible. Organizing things by project, area, review frequency etc is in fact very stable for me, so this is what I do, personally. Time planning is not very stable for me at all, so I try to reduce that to the barest minimum in order to reduce aggravation, confusion and work.
I totally understand that some people come from the opposite angle (especially people outside the GTD camp, but even from within the GTD camp). They feel that anything that lacks a date is "out of control" and they tend to have dates on a large number of items even if these dates are often quite "artificial" (a result of planning rather than actual facts about the action - they could reschedule the action with just a stroke of a pen without asking anyone, but they do not seem to be bothered by this.) For me, I like to see clearly which dates are "hard reality" that I cannot tamper with easily; any other dates than that I have never had much use for, since I know that even if I had them I could just reschedule them; they carry no significance and just blur my view. We probably all want something we can trust that feels stable. Some find stibility in a timeline; some find stability in other kinds of "task characteristics".