There are several layers to Alsacs question, and some of those layers have been addressed nicely in this thread. What I would like to add is that, though our offensive design can occasionally be quite effective, it often doesn't take full advantage of elements that could make it even better.
Before I expand on that point, let me first say that I am only looking at our ball-screen offense. We obviously use other plays, such as our pin-down screens and our "Double Fist" play. However, analysis of those are for a different thread. I also want to add that I admire Coach Boeheim's basketball mind. He is one of the best in the business for a reason.
Some of the posters here suggested that the answer to Alsacs question lies in personnel. Coach Boeheim's comments about limiting the offense this year based on the necessity of playing G at the point-guard spot suggest this is part of the issue this season. As some have stated, though, we have used non-shooting 4's and 5's to set high ball screens for years, often with little flow or effectiveness. That is bigger than a one-year personnel issue; it implies deeper design choices creating the problems that are sometimes seen in the half-court offense.
It was mentioned in this thread that Boeheim likes to keep his offense simple. True. He states so in his Syracuse Transition and Early Offense DVD. That said, though, he has actually strayed away from some of the concepts that he called "simple" in that DVD over the last decade. Though he always expected players to use ball screens when the options of a called play broke down, Boeheim's offense relies even more on ball screens now than it did in 2005. In essence, this philosophical shift makes our offense even more simple.
If it is more simple, why do we struggle so often to run an effective half-court offense? Part of the answer is that players often fail to execute basic fundamentals. Our screens are often not set properly when the ball handler begins his use of the screen. This helps the defense, and it can lead to illegal screens. Many times, our players don't dribble off the screen aggressively, looking to be the first scoring option, defeating the purpose. Additionally, they don't always take at least two dribbles off the screen to create space before deciding on the best play. Along with that, players often choose to refuse the screen when they shouldn't, which can let defenders off the hook, as does using the screen when it should be refused. Our ball handlers don't split traps or use a back dribble against traps to create space. Our bigs don't routinely slip the screen when their defenders hedge too hard. In fact, our lack of executing when the defense blitzes the screen is one of our biggest year-to-year issues.
So the issue is personnel, right? Get players with these fundamental skills and all will be perfect? Not quite.
Part of the issue is in our ball-screen design. We almost always run a single ball screen and, if nothing comes of it, a pass is made, and then a another ball screen is made for a different player. That's if we don't just create an iso for a player. Oftentimes, our spacing is worse on the second ball screen than the first. In short, this plan of attack is not difficult for defenders to stop. As long as each man defending the screen does his job, the defense wins.
Every coach I have ever seen discuss a ball-screen offense says the key is to incorporate multiple screens. For example, set a ball screen, and then set a back screen for the screener. If the screener's defender wants to jump out hard to trap, he is now screened far away from his man. When was the last time SU executed this type of screen-the-screener action in a high ball screen?
Another example is the ball screen/flare screen combo. The ball screener is a shooter (Cooney?). As he finishes his screen, another screener sets a flare screen for him to use. This can open up a three-point shot with the shooter squared to the hoop (assuming the player executes proper footwork). We actually used something similar to this in the first possession of the Dayton game. I almost fell out of my chair watching it.
Even simple screen/re-screen action can open up opportunities. Using more ball screens in transition could help. The key is make the defenders have to fight through multiple screening actions in succession. This doesn't allow teams to just stop a screen and win the battle. This is not a personnel issue--this is a design issue. Boeheim's simplifying the offense has made us easier to defend in some regards.
Back to discussing the non-shooter versus the shooter setting the screen. Theoretically, you'd want to be able to use both and, if memory serves, we have used both to a greater or lesser extent this season. However, both uses would likely be more effective if they weren't just isolated, random ball screens, but were more smoothly integrated into a series of screening actions. I'd like to see planned/designed ball screens from non-shooting bigs, shooting bigs, and shooting wings, combined with other screening action to set up ball screens or after ball screens, but that is just me. I love a good ball screen mixed seamlessly into other motion concepts.
That said, players have to learn how to read the defense and execute proper fundamentals for any offense to work. These fundamental skill issues and offensive design "flaws" are interrelated, and they have led to some of the offensive struggles we have seen as fans.