Re: [ng-spice-devel] personal note on ACS and NG-SPICE (fwd)


To ng-spice-devel@ieee.ing.uniroma1.it
From Alan Gillespie <alan.gillespie@analog.com>
Date Mon, 11 Sep 2000 15:57:12 -0400
Delivered-To mailing list ng-spice-devel@ieee.ing.uniroma1.it
Mailing-List contact ng-spice-devel-help@ieee.ing.uniroma1.it; run by ezmlm
References <Pine.LNX.3.96.1000911210312.2649A-100000@ieee.ing.uniroma1.it >
Reply-To ng-spice-devel@ieee.ing.uniroma1.it
Sender agillesp@epc.co.uk

Some comments .....

> There are good reasons why I did not base my research on Spice.
> There are also good reasons for the move to C++.  I only wish I had
> done it sooner.  The design of Spice is such that if you really want
> to move ahead, it really will require a full rewrite.  Spice-3 is as
> far behind today as Spice-2 was when Tom Quarles undertook the job of
> converting it to C.

I'd be interested in knowing what's different about ACS. I did download
it a while back, and I quite liked some of the stuff I read in the
documentation.
It seemed to me, though, like it was just doing the same thing as Spice,
i.e. solving a matrix of linearized current sources and conductances.

So what exactly do you mean when you say "as far behind" ? Do you mean
just in terms of the language it's programmed in ? Or the structure of the

program ? Or the way it actually performs a simulation ? Or what ? I'm
not being defensive here, I'm genuinly interested.


>
> This all makes me wonder why it is so important to base it on Spice,
> especially considering that none of the group has any real
> personal attachment to it.

I have been hacking the Spice code for the last couple of years, so,
like it or loath it, I do have some kind of an attachment to it. But I
would
be willing to give that up if there's a better solution that's easier to
learn,
and maintain.


>
> With all this in mind, I would like to offer you ACS.
>

Wow, that's quite an offer.

>
> The biggest down side I see with this is that there are some in the
> group that will not touch C++.  This seems strange to me, because I
> found C++ to be a big productivity improvement over C.  Much more
> than the improvement of C over Fortran.  But, to each his own.  If
> the system is modular enough it doesn't matter what language the
> pieces use, and C++ modules can be C inside.
>

I, for one, really hate C. To me, C is the BASIC of "structured
programming".
C++ is definitely a step forward from C, although I haven't used it much.
I
only use C 'cos I have to. I still prefer Pascal (although it is a pain
for a few
things).

Since I've only recently joined, I haven't seen any anti-C++ messages
yet. Speaking of which, am I becoming the noisiest one here ? Or am I
missing some messages on other lists ?

Anyway, while it would be nice to respond enthiusiastically to such
a generous offer, it looks to me like you guy's have put a fair bit of
work
into restructuring Spice. How long would we need to evaluate ACS to
see whether it really gets us further on than we are ? How "credible"
as a re-vamped ACS versus a Spice rework ? I don't mean to be
cheeky, Al, it's just a matter of perceptions, i.e. Spice is an industry
standard, a benchmark.

I remember noting on the ng-spice web page that you were really
stating two goals - one was improving Spice, and the other was a
completely new rewrite. Is the ng-spice-rework supposed to be
the "all new" creature, or are we still improving Spice ? Could ACS
be the basis of the "all new" version, while the rework is just a "fix"
of Spice ? Do we have the resources to both ?

Cheers,

Alan



Partial thread listing: