[sdiy] midi clock

[Magnus Danielson]

From: patchell <patchell at silcom.com>
Subject: Re: [sdiy] midi clock
Date: Mon, 02 Dec 2002 17:27:39 -0800

>     Beleive it or not, I do that all the time.  The Xilinx Field
> Applications Engineer was amazed when he saw the designs I did without
> simulation.  When the place I worked for originally started doing the
> FPGAs, we could not afford the simulator (it was an extra $4000 on top
> of the $10000 for the compiler....the software is much cheaper now...ie
> FREE).  I got used to doing it blind...and for the most part, it was
> pretty easy to get the designs to work the first time, or at least
> nearly so.

It all cooks down to how you do the design work. IF you've got part of it into
routine and care about the other details, it's pretty easy to do rather larg
designs more or less flawless without a simulation. The actual gatecount isn't
necessaryilly a good measure of complexity. If the design has a large
selfsimilarity, such as a wide DSP dataflow gives, then the complexity of the
design is much lower. My experience is that the things that mostly caused
troubles are exactly the points where I didn't spend sufficient of analysis and
planing, or more bluntly, care, into my designwork.

The same thing really goes for software. For instance, I once did a 1000 liner
from scratch, first compile round showed 3 typos (that was 3 hand-on-forehead
typos, i.e. I agreed with the compilator once I saw the offending code), and
the running code I'd reversed one thing (which existed in both forward and
reverse order). Other than those things, it ran straight out of the box. For
that 1k liner I didn't actually do any design-work other than what was done
while coding the implementation.

So, it can be done. For those that really want to see a story about it, go and
search for the Multics webpages and read the "It can be done" story. I'm sure
you all are running Multics on your 635s out of your bedrooms, arn't you?

PS. Anyone in for some Software Engineering information, get and then read
"The Mythical Man-month" which is a good read and learn thing while not pushing
this or that "work-method" on you... it pushes many on you, just for the sake
of crushing them. Say one flawed way of thinking per chapter. Many chapters.
Got to read it again...

Cheers,
Magnus