I am currently modeling several kinesthetic tasks, which has traditionally been my weakest system. I am making steady improvement in my ability to "stop the world", though I still have the ghost of the echo of the whisper of my internal dialog. One skill is piano playing, and the other is ballroom (swing/salsa/etc.) dancing.  The problem that I'm finding in both skills is that I am unable to perform the skills well without an external visual cue to "remind" my K system the patterning. In other words, my kinesthetic memory is weak. Once I have visually seen the sheet music or the dance pattern, then I am reminded and am able to perform. Any suggestions on building this piece of the skill? Also, should I focus more now on a completely silent "stop the world", or continue modeling and trust it will soon fade completely? On another note, I am also finding that I am modeling most effectively when I am modeling one skill set at a time, and allowing that skill to float in and out of 1st and 2nd attention all day long. Standing in the grocery store tonight, the music caught my legs' attention and I began the micro muscle movements of the cha cha. When I was modeling software engineers, the most elegant exemplars said that they had thier software design problems in thier head all day long, all night long, and even over the weekend, and would get flashes of insight in the strangest places. Has anyone else found anything like this?