Wow. Thanks for the mention in the article. Here's something I thought of after I hit Submit the last time.
The launch has two phases. In the balloon ascent phase you are mostly worried about spin and wind buffeting the conjoined craft. In the launch phase I gather that you want the rocket plane to point up to gain additional altitude. That's smart, because you won't get any lift from those small wings until the craft is back down to ~10km. Anyway, there's a pretty lightweight solution to this: Mount an R/C servo in the beam.
If you shorten the back lines so the beam is almost horizontal, then the wings and body present less surface area to the wind. You would even get some lift from the wings if you put a streamer tail on the beam to point everything into the wind. Once you get above the jet stream, then unwind the rear lines to lower the beam into launch position for the rocket plane.
This can be accomplished many ways. Most of them are fairly simple and foolproof. A straight timer solution would only take a few 556 chips. One to generate the PWM signals (10/90 for ascent, 90/10 for launch) to the servo. Another to trigger the first one into switchover after ~45 minutes.
Servos are also easy to modify for continuous winding if you need more than 270 degrees of rotation to get into position. Then just tie a bead to the end of the line and have it drop off the pulley so it doesn't wind back up.
Also, my aviation experience is limited, but, as far as it goes, ice buildup is a problem when flying through thunderheads (rain or snow clouds). If you are launching on a clear day, then any ice would sublimate as fast as it lands. Private planes fly all day long in clear skies without wing heaters.