If i have DOS software i can reverse engineer how the
hardware is driven and build my own hardware.
which may have different transistors and thus give different results.
one problem with this is that you have no means of finding out if the cells are fully programmed by your kitchen table programmer.

I have been informed that one of the reasons that manufacturers do not release parallel programming info any more is that many kitchen table programmers did not fully charge the flash cells. This led to chips that "tested" (I hate the reliance so many put on that) beautifully but after a short while lost bits of the program. Of course the kitchen table programmer users came down on the manufacturers "my tested chip is failing" totally ignoring the fact that the fault was incomplete (dis)charge of the flash.

So, go ahead, reverse engineer and make your kitchen table programmer, I wish you luck, but DO NOT come back and whine when your "tested" chips loose the program after a year or two because of something you "overlooked".

Erik