Documentation is an issue indeed. I don't think that you really need the inner information about NAND to do the original and innovative research. It really dep

Documentation is an issue indeed. I don't think that you really need the inner information about NAND to do the original and innovative research. It really depends on the area you want to work on, the SSD firmware level might require that but the SSD makers are already on that route (some better than others). The other level is to not pay too much attention to the differences between NAND chips and just implement something at a higher level to push the hardware-agnostic smarts to the OS. The OpenSSD also lacks documentation, last time I looked at it there was no info on how to do NCQ on the SATA interface, without which there is no talking about a speedy SSD.