1875 made an amazing attempt this year with an idea one of their students had to do something like this. They found a way to do a version of what you are talking about for super cheap using a boat antenna ($15) some simple CNC'd X brackets to run cables up the antenna with, and mounts at the base. At the Florida Regional they were having problems staying within the starting dimensions (they would bend it around to look almost exactly like a question mark to start with and then it would spring out at the start of the match) as well as controlling it (they had one person controlling the 4 motors himself if I'm not mistaken, so considering they had no sensor feedback it was really, really, really innovative and definitely something I think should be pursued more). For the "complex" movements it achieved (360 degree turret arm that could extend out 8 feet or so from the robot) with such simplicity (4 cables and a boat antenna) it was quite a nice development within FIRST. I know a big issue with it was the spring force of the antenna is what was used to naturally make the arm go back up as the 4 other cables were released, so it was fine to pick up something like a tube, but to pick up a Tetra would probably be impossible (at least with a boat antenna) just since the weight is so massive in comparison to those tubes.
Here are some pre-ship photos of their robot, I wish I had some from the competition, but like I said they did have some problems with it and you can get the basic idea from these... very cool stuff.
http://www.chiefdelphi.com/media/photos/27108
http://www.chiefdelphi.com/media/photos/27096
http://www.chiefdelphi.com/media/photos/27078