On Fri, 2011-10-14 at 15:01 -0400, Dave Tweed wrote:
> I apologize in advance -- I'm really not trying to sound elitist here -- =
but
> if you need "many" iterations to debug a simple design (your word, not mi=
ne),

A "simple" design in the professional world rarely is.

For example, yes, this chip is "simple". But chances are you are going
to have it on a board with MANY other parts, designed by MANY other
individuals, depending on MANY docs, with likely MANY architectural
changes (been there, done that).

If your design is "one man" and you are the customer then I can agree
with you, if you spend the time to do it perfectly on the onset chances
are you won't need many iterations.

In the real world, in my experience, rarely is a design one man, and
rarely does your customer know what they want. Chances are the specs
will change. Other then changes, chances are the specs will be
interpreted by one design team one way, and by another another way (I've
seen it many times, in one case the docs specified a protocol in such a
way that it would be easy to interpret how the chip worked in two
completely different and incompatible ways, one design team interpreted
it one way, the other the other way, in the end, the design would have
been deadlocked had the mistake not been caught).

> then what you are doing is not engineering. I have done and have seen oth=
ers
> do very complex designs, including FPGAs with tens of thousands of gates =
and
> large PCBs with hundreds of ICs, in just 1-3 iterations.

Actually, it's VERY engineering, here's why:

With boards, I agree with you (I usually work on designs with many FPGAs
on a board, so millions of gates is the result). The point is it's VERY
hard to change a board after it's fabbed. So a larger amount of
engineering effort is put into the board so that there are zero errors.

It's a tradeoff. If your design CAN be changed, then getting it in a
physical form is accelerated, since problems can be fixed after the
fact. If your design is hard to change (i.e. a OTP part, or a PCB) then
ALOT more up front engineering time will be spent to ensure everything
is correct.

That is engineering: weighing the pros and cons and making a
compromise.=20

For me, adding a OTP part to a board is not worth it, given the
thousands of alternatives that don't have that limitation. I understand
others might find this part useful, for me, no thanks.

TTYL


--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.