Here is all the documentation you really need. The reference is Boyd Schierle and Beckwith, 1988. (There is a link to the abstract on the left.) The maxH value is the tallest peak in the hydropathy plot with a window of 19 residues and the JTT2 scale. P(int) is the probability that the protein would be integral if it were an E. coli protein. All other reasonable scales and windows give similar results. If the P(int) is near 1 or near 0 you have a good indication about the probability of your protein being integral in E. coli and a suggestion for other organisms. If it is between 0.1 and 0.9 the maxH value is not much help unless you have an E. coli protein or a protein from a similar organism. We think the maxH value may be a basic property of all proteins that efficiently targets those with high values the membrane. This basic mechanism may be overridden in a simple way in Bacillus and similar organisms where some signal peptidases may recognize proteins with signal sequences which have long hydrophobic cores and, more subtly, in eukaryotes where some mitochondrial inner membrane proteins coded in the nucleus may require low maxH values in addition to mitochondrial import signals for efficient localization.