Hi John, I too noticed that Miller had been "artfully vague" in his definitions regarding bits/chunks, and perhaps my question should have been whether or not you were aware of any research that sought to better explicate this "code" and/or the definitions of these "bits/chucks"? (I may have been getting ahead of myself asking the next questions about how our strategies to attend to those chucks might affect "state".) Your example served well to highlight the differences, and subsequent difficulties in accurately quantifying the bits being consciously attended to, although in Miller's work it does seem that sorting into digital categories may be a first step in defining a "bit/chuck". (e.g. Visual only: large vs. small,  round vs. square, blue vs. red, light vs. dark, etc.) My thought was that using this a base from which to start, it seems that the difference from being able to sort 32 (2epx.5) vs. 512 (2exp.9) types of "things" consciously would represent a significant difference in performance. You said, "Thus until the code itself is fixed, specified, the question is not well formed." A point well taken. I also find this fascinating, and have ordered a copy of Babel-17 from my local Si-Fi bookstore and thanks for the suggestion. I'm always looking for good Si-Fi). ;-)