[sdiy] Re: Bottom Ten ICs

[Rainer Buchty]

>Which part of 'I still think the RAM approach was a crappy design
>decision with no futureproofing and hey - the rest of the world seemed
>to think so too' was I not being clear about?

Register windows existed (and exist) within many processors.

Again: for the time of creation it was a sensible move to put that 
register bank entirely into RAM, because putting that much registers 
on-chip would have made the CPU overly expensive.

*Today* (well, make that "since mid/late 90s") we have an entirely 
different situation with processors running one and a half magnitude 
faster than available RAM. This is why you don't see this approach 
anymore.

Besides: any architecture offering indexed-indirect and indirect-indexed 
addressing modes sort of mistreats memory as some sort of register 
resource. Very famous on the 6502, btw. The Zero page gave you 128 
16-bit pointer registers.

>Too many engineers believe a great piece of hardware will sell itself.
>What companies like Intel and MS proved was that your product can be a
>steaming pile of crap, but if you have the right market leverage it
>will sell in huge quantities anyway.

Works for other areas as well. VHS anyone?

>I'm not saying I'm happy about this, but I am realistic about it.

And this is quite sad. Marketing will do anything but innovate. What 
marketing does can be seen on the music market: mainstream, mainstream, 
mainstream. No experiments. Experiments are costly, and costs are 
something to avoid under all circumstances -- especially if the company 
is run from quarter to quarter but not with a longer-term perspective.

I like the way Scott Adams phrases it: if the past were entirely 
marketing driven we would have not invented wheels yet, cause marketing 
would have convinced everyone of the superb braking qualities of 
trapezoids.

And this "fire" thing... If it's such a great thing why didn't someone 
else invent it already?

Rainer (who has worked as FAE and FSE, and happily returned to good old 
	research afterwards)