One thing they all have in common is that none have set the world on fire. I think there are two reasons for this limited success, and both are about to change.
Reason one is that the economics of solo driving have always been relatively reasonable in the U.S. That could change as the economy continues its downward spiral. People will be looking to cut costs anywhere they can, and they will give up flexibility to do it. That's new, or potentially new. And in developing countries the economics of single passenger autos is less favorable. People will have iPhones long before they have their own cars.
The second obstacle to ridesharing is a sense of control. Imagine finding a ride match on your computer then walking to the sidewalk and hoping it actually shows up on time. Or imagine walking to some central pickup location and hoping there are enough drivers for the number of riders. You would feel you had no control. That's a stopper. But I can imagine a certain type of iPhone-like application that could give you back the feeling of control. I will explain.
First, it would help a lot if you could easily negotiate a ride from the iPhone as opposed to needing a computer. That helps if you need to make a quick change in plans. That's the first part of giving you a sense of control.
Next, the application should use GPS to draw a map of your location, with blips for the cars available for ridesharing. You select the nearest blip and a bio comes up telling you something about the driver, including his primary profession, age, a photo, and a picture of the car. If you don't like something about that potential ride, move on to the next nearest blip. Again, you have a sense of control. Likewise, the driver could reject you as a passenger after seeing your bio.
After you select your driver, and he accepts, you can monitor his progress toward your location by the moving blip on your iPhone. As with the progress bar on your computer, the feedback will give you a sense of control. And with an iPhone you can stay entertained while you wait. That helps make the time go by, and again gives you a sense of control.
I also imagine that all drivers would have to pass some sort of "friend of a friend" test, in the Facebook sense. In other words, you can only be a registered rideshare driver if other registered drivers have recommended you. Drivers would be rated by passengers after each ride, again by iPhone, so every network of friends would carry a combined rating. That would keep the good drivers from recommending bad drivers because the bad rating would be included in their own network of friends average. That system needs more thought, but you can see where I'm going on that. And the same system could be applied to potential passengers. As the system grew, you could often find a ride with a friend of a friend. And that automatically gives you something to talk about too.
The big fear people might have is that strangers would commit crimes against them. But remember that the system would have a record of every ride matched, including the identities of the participants, and a GPS record of where they were and when. A rideshare car would become the very worst place for a criminal to commit a crime.
Apple could make it happen just by good design and of course the coolness factor. The profit potential is huge, for both the system operator and drivers, so that imparts some inevitability to this idea. The U.S. will have too many legal barriers to be the leader in this sort of thing, so I expect it to catch on in other countries first. Once proven elsewhere, the U.S. might take a look.