I'm renovating a bathroom which includes retiling the bathroom floor. There is nothing wrong with the existing porcelain tile except for the colour and texture. It is set in a 3/8" bed of regular unmodifed thinset. I had stapled diamond stretch mesh metal lath to the 3/4" OSB subfloor and laid electric heating cable into the mortar. There are two schools of thought WRT to modifed versus unmodified and then there is the crack isolation membrane debate as well. Modified is usually called for over plywood or OSB. One school of thought is the wood wicks out the water from the thinset reducing the ability for it to wet cure resulting in a weaker bond. The other school says that the latex in modified thinset is a bonding agent. In reality I think both are true, the latex is a bonding agent and the latex is also an anti-dessicant slowing down the loss of water. For some reason that I don't fully understand, porcelain tile manufacturers call for modified thinset. Porcelain tile does not wick away water so that blows that school of thought. The DITRA/KIRDI makers say DON'T use modified for bonding sealed surfaces. Both the DITRA and the porcelain tile are sealed meaning they don't absorb water so all the water in the thinset has to leave via the grout joints. They say the latex in modified thinset slows down the loss of water. DITRA or cement backer board is usually prescribed for crack isolation. I did not want too good of a bond to the OSB so that if there was movement due to humidity changes, the mortar bond would give and the metal diamond lath would prevent cracks telegraphing to the tile or grout. It essentially was a trowel-in-place cement backer board and has worked well as there are no cracks in more than 11 years of use. The problem now though, is that I cannot chisel up the old porcelain tile without taking the mortar with it, damaging the heating cable in the process. I tested it by removing a tile from under the old shower base that I removed as there is no heating cable there to damage. The unmodified mortar is not releasing from the underside of the porcelain tile but it is releasing from the OSB taking up the diamond metal lath with it. My plan now is to lay new porcelain tile over top of the exsisting porelain tile. The old porcelain tile has a slate look so the face of it is textured which should help with the bond. Most of my search results say to use modified but that goes against what they say about DITRA. To further complicate this, I will be using epoxy grout so the water is less likely to leave through the grout. That means I need to wait longer before grouting so that means I have to go without my shower. Our other bathroom has a soaker tub but no shower. I hate bathing. There is one other school of thought, to use a whitewash of latex and water. When bonding new concrete to old, the old concrete gets a coat of watered down latex first. This is also prescribed for thinset mortar over plywood or OSB. The latex whitewash wets the wood so that the wood doesn't wick away as much water from the thinset and it also acts as a bonding agent. Being that both wood and concrete are porous, the water would not be trapped as it is on porcelain. Thoughts?