... Clearly a tragic number of incompetent people were put together onto this team. I've often made the observation that teams often cosist of six or more people are given a task, and two do the work, while the remainder fight over who will take credit. Nobody is charged with ensuring quality or even functionality until it's too late. Often the most junior, often the least competent, is charged with that.

That's evidence of bungled management from above. While there are many people involved in manufacturing and in marketing, not so many are engaged in development and testing. When a product makes it into the field, while totally non-functionaly in nearly every respect, it's not an "error." It's total breakdown of management discipline. This device was on a production-scale PCB with an on-chip DSP as well as the ARM7TDMI core, a boot-loader in flash, a High-Speed codec pair, operating at a rate they even admitted didn't belong on this board, an SDRAM channel, and support software for a LINUX implementation. Nearly nothing of that worked! No bootloader, no on-chip flash, no SDRAM ... as a result, they produced an SRAM board too small to host a LINUX implementation, making their entire development suite, which relied on LINUX, useless.

Clearly, management had gone to sleep. Even worse than that was the practice of continuing to pitch their no-longer-useable eval hardware as top-quality stuff that would support the evaluator in booting an OS by means of which he could produce end-product firmware and software. This only weeks before discontinuing delivery of the chip. Further, they discontinued the chip before notifying active customers.

It was my understanding that those several (more than the three with whom I spoke) individuals were, at my last contact with ATMEL's local representative, who, BTW, subsequently resigned, were still with ATMEL and likely to remain so. That suggests that ATMEL finds bungling of this sort totally acceptable.

THAT is why nobody I know uses ATMEL products of any sort.

RE