1) Heat as a measure for the energy a photon is carrying is OK, but photons should not be able to exchange that heat with each other, as collisions between photons are not possible. Heat exchange between photons and matter is OK, but I would suggest to:
2) Absorb a photon when heat is exchanged. If one gives elements different probabilities to reflect (or even transmit in the case of glass) a photon instead of absorbing it, the characteristics of materials can be pretty different (broken metal would in most cases absorb a photon, while solid metal reflects it as a default etc.)
You are correct about how the photons should behave when transferring energy to a material, but, photons can and do indeed collide with each other, just, typically, there is no noticeable energy transfer between colliding photons because, unlike other particles of mass, their energy is related to their wavelength and not their mass/velocity. The reason, therefore, that we don't observe photons transferring thermal energy between each other is that their energy is carried in a form that is only transferrable to mass other than photons and not because photons can't collide. This is also due to the fact that photons transfer energy to objects by being absorbed (as you stated) and photons can't absorb other photons.
EDIT: In other words, I agree with 2.) completely and I agree with 1.) as well, just for different reasons.
Excimer-Sun-Software I have a physics background myself, wanted to keep it easy In TPT-terms the photons would be not-colliding.
The heat exchange between photons does not make much sense.
Just one example: A photon cannon, shooting 'hot' photons at a metal surface would deliver a short volley of hot photons, and after that, the reflected photons would cool down the incoming ones. Not much of a photon, right?
It is not about realism, but why call something photon if it doesn't even compare to the real ones? And the no-rest-mass characteristic is probably the core characteristic of a photon.
I have no problem with photons colliding with eachother as long as they don't transfer thermal energy between eachother in the process. I don't think that there's a good way to keep photons from colliding without seriously slowing down PT. I can write code to make photons be absorbed upon heating things, and to make photons heat everything except for other photons, without slowing PT down measureably (in theory, since I haven't tried yet) but I just can't think of a good way to write code to keep photons from colliding without considerably slowing PT.
Stack them. We have 8 general directions of photon movement, photons can be packed into 8 streams. A stream can only receive an additional photon (from clone, LED etc.) if there is space in that particular stream. Basically, a dx=1,dy=1 photon would be in a different stream then a dx=0,dy=-1 particle and therefore could not collide.
The worst result would be a total of 8 photons per pixel (if you really manage to fill all area with photons of all directions)
I can write code to make photons be absorbed upon heating things, and to make photons heat everything except for other photons, without slowing PT down measureably