If they made it clear, ABSOLUTELY clear, and complete, how one is supposed to program each of their parts, particularly such that someone, anyone, smart enough to follow the directions, yet dumb enough to want to do so, could do it, this would never be an issue. Sadly, they (ATMEL) are not alone.

They're not the only company that closely guards their parallel programming requirements, though that no longer helps them.

If I were going to program mfg XYZ's parts, or, for that matter, mfg ABC's parts, I'd prefer to have a parallel programmer rather than a serial one. However, if they want to promote ISP as a valid programming scheme, they should publish precise specifications for that, too, including all the timing and voltage level details.

When there's an ATMEL-compatible (or NXP-compatible, or anyone else, for that matter) programmer, it's likely someone, somewhere will attempt to use it on the wrong IC. The result could range from failure to function to complete and total destruction of the target IC.

There's a problem here, but it's being ignored, not only by the chip makers, but by the user commuity as well.

RE