Must be infectious "good enough" mentality from CPQ heads:
...ibm products and componenets are subjected to a thorough battery of performance tests and quality checkpoints that identify potential flaws in memory and other components during our design phase before shipment to customers. Questionable parts are disqualified from shipping in an IBM product...
"someone in HP ordering the wrong part ..."
Those damn Indians have problems reading English. You need to give them an incentive - try a bowl of curry and a cow next time. Then they will get your order correct.
HPQ Knows Indians
No, it means that either IBM hasn't found it yet or they didn't implement the functionality that leads to the failure.
It is a very far reach to assume IBM caught it. You are implying IBM knew about the problem for almost 2 years and didn't inform either Intel or the memory vendors. I'm having visions of IBM having problems getting parts.
"This memory insue echos the same."
I agree, most likely some bean-counting purchasing agent decided "I could save a bunch of money by substituting this part! It has the same number of leads, so it will work." Same with the RAMDAC fiasco.
Yes, I was just making an educated guess on why swapping memory modules fixed the problem. HP is placing the blame on the memory vendors, not the lack of qualification by HP. That doesn't sound very likely.
Recall that the new HP (AC but before +cpq) had to recall a bunch of Jornados because the idiots ordered 8 bits instead of 16 bit color modules.
No body caught the problem until customer's complain. That was a manufacturing flaw that escaped all the checks, or lack of, from new product qualification, documentation, final test to OQA.
This memory insue echos the same.
Implementation problem or incorrect part? HP seems to be fixing it by installing new memory modules. So, are they substituting faster memory for what it was shipped with? That would explain it, but it wouldn't be the memory makers fault in that case, would it?