Reasons have been found for thinking that the minimum diameter of channels of a given length to support superconductivity at room temperature through films of oxidised atactic polypropylene (OAPP) is considerably larger than found in a model for Bose condensation in an array of nanofilaments [D.M. Eagles, Phil. Mag. 85, 1931] used previously. This model was introduced to interpret experimental results dating from 1988 on OAPP. The channels are thought to be of larger diameter than believed before because, for an N-S-N system where the superconductor consists of an array of single-walled carbon nanotubes, the resistance, for good contacts, is R Q /2N , where N is the number of nanotubes and R Q = 12.9 kΩ [See e.g. M. Ferrier et al., Solid State Commun. 131, 615 (2004)]. We assume this would be 2R Q /N for a triplet superconductor with all spins in the same direction and no orbital degeneracy, which may be the case for nanofilaments in OAPP. Hence one may infer a minimum number of filaments for a given resistance. In the present model, the E(K) curve for the bosons is taken to be of a Bogoliubov form, but with a less steep initial linear term in the dispersion at T c than occurs at low T . This form is different from the simple linear plus quadratic dispersion, with a steeper initial slope, used in my 2005 paper. A combination of theory and experimental data has been used to find approximate constraints on parameters appearing in the theory.