Modern Electronics (was Re: List charter mods & headcount... ; -))

From: William Donzelli <aw288_at_osfn.org>
Date: Sun Jun 20 00:35:27 2004

> Tools are great, I use them myself. But you have to have some knowledge of
> what's going on to know when the answer the tool gives you is wrong (maybe
> you didn't feed the tool enough information, forgot something, or there is a
> bug in the program, etc).

Obviously you have to know *something*. I wouldn't expect an
accountant to make a video card, nor an engineer to do an estate plan.

> Lots of solutions that look good on a computer end
> up not working great in the field because of many things that are not
> modeled in your tool.

I am not arguing this point. I have plenty of experience on the
manufacturing floor to know all about this (a classic example is
mechanical interference of parts on a circuit assembly).

I should say that modelling is getting much better all of the time, and
there may be a day when "the real world" is simulated very nicely inside a
model.

> (that you DO use believe it or not for the rest of your life)

Yes, some you use (I did say "*basics*"), but many, no. Look at something
like Transistors 101. Most engineers never use all of the small signal
math - when they need a transistor, they just pick one of of a book, and
more often than not, just use a canned circuit (maybe even the one in the
book!). Same with computer architecture - engineers pick microprocessors
out of a databook or macros out of a library, but they don't sit down and
start thinking about how many registers to have, and how the microcode
will work, and how many function the ALU will have, and an endless list of
other variables one has to deal with when designing a microproccessor..

> Specialization is ok if you like doing just one thing and don't mind being
> out of a profession when your specialty is obsolete (especially in any
> programming or electronics field).

Lets not go overboard here. Engineers are supposed to change. Do not
forget that in the electronics field, an engineer that does not keep up
with the times is a dead engineer.

However, if you are specialized in a field, and you need to change jobs,
specialization is a very good thing to have on a resume if the
specialization fits the job. "Jack of all trades" engineers that are
really good at everything are pretty rare - always have been. Engineering
managers and HR people know this, so being too versatile can be a
liability.

> Tools just do the time consuming calculations, you still
> have to tell them step by step what to do so I would not say they do all the
> thinking.

Well, not all the thinking (maybe a bad word on my part). For example,
these days with all the HDLs out there, you tell two variables to add
themselves, rather than design an adder. Gate level design is a dying,
almost dead, profession, reserved for relatively few (read:
mega-optimized) applications.

In reference to the "step by step" aspect, those steps are getting pretty
wide, and are getting wider all of the time. I think this is the root of
this whole discussion - people are lamenting the days when the steps were
tiny, or did not exist at all, so an engineer had to know all aspects of a
design. This is simply not the case in this age. Teaching kids to take
little steps will only result in engineering projects that never get
completed in the real world.

William Donzelli
aw288_at_osfn.org
Received on Sun Jun 20 2004 - 00:35:27 BST