Mees, in his Theory of the Photographic Process, describes this process very briefly. Only four pages out of one thousand are dedicated to it.
He basically states that:
1. The exposure must be approximately five times higher than in the case of chemical development. With high speed films it must be increased even more than that, so much more that it would generate a lot of fog. As this is the 1942 edition, I suspect that by high speed films he means ASA 200, or at most 400.
2. Developing after fixation is preferable, because fixation will also reveal those latent image nuclei formed inside the grains, which are not accessible beforehand.
3. Acidic fixers must be avoided, because they can dissolve the delicate latent image. Although he doesn't mention them directly, I assume that rapid fixers are totally out of the question for the same reason.
4. Even basic solutions of sodium thiosulfate can destroy the latent image if they are aerated.
5. The safest fixing agent is sodium sulfite, because it will never attack the latent image. However, sulfite cannot fix films with high iodide content.
That's all he says about it. Hundreds of pages are dedicated to chemical development, and to developing agents, and only four to physical development.
What I have is the 1942 edition. Maybe newer editions are more detailed; I don't know about that...