Re: [Gnucap-devel] SPICE to gnucap transition

[al davis]

On Monday 19 March 2007 12:38, Scott Dattalo wrote:
> al davis wrote:
> > There is always the question of what to do first, what to
> > hope someone else will do, etc.  If I did all that, there
> > would be no time left for the good stuff, leaving no reason
> > for gnucap to exist at all.
>
> I've been aware of gnucap for more than 5 years. As far as I
> can tell, you are the only one that uses it! Maybe there are
> others... but they're very, very quiet.

There are fewer users now than 5 years ago.  There are a fair 
number of people who follow it, who are well known as experts.  
They do tend to be quiet.

The fact that there are few users is further evidence that the 
best plan for the future is to break completely from spice, 
evaluate where we should be, and go there.

> <snip>
>
> > There is a serious need for beginner documentation.  I plan
> > to do that, when Verilog works.  For now, it doesn't bother
> > me to lose the people who want Spice as Spice.  I am glad
> > that ng-spice exists, to take up this slack.
>
> Wow! Don't you realize that people don't want SPICE for
> SPICE? They want a circuit simulator for analyzing circuits
> and SPICE is the only open source package capable of doing
> it.

Of course I realize that people don't want SPICE for SPICE.

Most of the last 5 years, I was in a job situation that was very 
damaging.  About a year ago I decided to take care of that, and 
take another look at my future, and the future of gnucap.

If gnucap tries to look like spice, people will think of it as 
spice, and not see the benefits.  They will see the 
deficiencies.  They will see arbitrary differences as 
deficiencies, even if they are really improvements.

If I try to make gnucap act like spice, that will take all of my 
time, and the real goal of moving beyond spice will never 
happen.  Not only that, it will never be spice.

The future of simulation is not spice.  People complain about it 
all the time.  The future of gnucap is to push away from spice, 
totally, and address those needs.  The future is in the newer 
languages: Verilog-AMS and VHDL-AMS.  That will be the emphasis 
of gnucap.

gnucap will be a *AMS simulator that is most emphatically not 
spice.

> You need people like me who know SPICE's strengths and
> weakness and are willing to show how gnucap is better. In
> addition, I've got experience with simulators, C++, dynamic
> libraries, etc. and probably could help out in more ways than
> you can imagine. On the one hand you acknowledge you need
> this help but on the other you're pushing back. I'd really
> like to help out here!

I am pushing back from spice.  I am acknowledging that the 
gnucap you will see a year from now will be very different from 
the gnucap you see today.  I am also acknowledging that any 
effort to prop up the way it is now, at the expense of the way 
it will be, is not productive in the long run.  I would hate to 
have you do lots of work documenting the present, only to have 
it change as soon as you get it done.

Where you can help is to help with the vision of the future 
gnucap.  Given that the primary language will be 
Verilog-AMS ... now what should the commands be?  How should it 
interact with VHDL, Spice, gschem, PCB, etc.. ?  

Look at other simulators, particularly the ones I don't have 
access to.  Among them ... Spectre, Saber, Touchstone, ....  
What kind of command sequencing and scripting do they use?  
What do you like and dislike about them?  How consistent are 
they?

What commands should be available first, so Verilog-AMS can be 
used as an alternative to Spice, in a way that is better for 
beginners than Spice?

We have the opportunity to make a big impact, but breaking from 
Spice an absolute must.  The existing hybrid situation must go, 
as soon as possible.  Your experience is a perfect example of 
why!

Then, after breaking from Spice, now how do we handle the 
transition? 

On Monday 19 March 2007 12:38, Scott Dattalo wrote:
> So here's an offer:

I will answer that part separately.