Well, Dallas, in 2000-2001, released a datasheet with their parallel programming circuit in it, and the timing and command syntax. Frankly, I wish I had the old datasheet. I probably do, but don't know where I stashed it. I wish I could find it, just to try building a parallel programmer for the things, based on a small CPLD. Most devices can be programmed MUCH faster in parallel, and a small shop that wants to program 1000 pieces often doesn't want to pay to have the programming done by the distributor or manufacturer, as it doubles his cost. Those garage operators don't value their time at $25/hour as the Neanderthals at the distributor do.

Intel published the parallel programming spec's for their i875x parts, too, as they'd done with the i874x-series. If it had been such a problem, I imagine that they'd have left that information out of the i875x datasheets.

Since there were about 25 years' industry experience between these two events, it's possible you're seeing things incorrectly.

I fully agree that "kitchen-table" programmers don't deserve the discussion that they get here. If people want to build their own programmer, they should have sense enough to use the correct components and reasonable construction technique. I remember one complaint that a guy having built a kit (not in this forum) ended up with a non-working programmer. The picture he submitted along with his lament showed he'd been too "frugal" to buy and install a ZIF socket and too sloppy to ensure that his MCU's pins were all in their socket. Frankly, if someone tried to get me to do something, yet, on the way, left one of my legs hanging out the window, I'd probably be a mite uncooperative as well.

The main problem, of course, is with documentation. It's not with kitchen-table programmers, necessarily, either. It's often with the commercial programmers that are sold to program component ABCvNNN but that doesn't have instructions recommending against using it with component ABCvNNNa. ... and then, of course, there's the guy who says that his ABCvNNN programmer works on XYZvAAA, or the one who says he's had success with a different software driver.

I think it's unreasonable for any manufacturer to expect the hobbyist to buy his $100US programmer when he has to save up for a month to get his $4 MCU. That guy would proably do a good job building his kitchen-table programmer, if the mfg would just provide the necessary specifications and do so clearly, correctly, and completely. After all, a lot of high-volume applications have been born in the garage or basement laboratory.

RE