Re: gEDA-user: Why use gEDA?

[al davis <ad136@xxxxxxxxxxxxxxxx>]

On Sunday 25 February 2007 16:08, Peter Clifton wrote:
> I'll chat to my colleges here who use PSpice for all their
> simulations, and find out what it is (if anything) which
> keeps them using it rather than anything else.
1. Inertia
2. The books push it.
3. The professors push it.
4. Publishers insist on it.
4. The books and professors limit what they ask for to what 
PSPICE will do.

> I suspect, that in some cases, it is the ability to reproduce
> and use work which others have done using the PSpice
> specific models or extensions to the language.
> 
> (As a non-PSpice user, I don't know what those are exactly,
> but I know  I've found difficulty using SPICE code / models
> from the net with ngspice in the past).

Try it with HSPICE, Smart-Spice, Multi-sim, etc.  any other.
They are all incompatible.

That's why industry is moving to Verilog and VHDL.
That's why gnucap is moving to Verilog and VHDL.  
That's why gnucap is moving to having everything in plugins.

> I have spoken to an academic here who has used PSpice 
> (and other spices) since before PSpice had a GUI, and he
> maintains that PSpice  has the _best_ solver of all the
> packages available. 

I doubt if he looked very hard.  People adapt to whatever they 
are using.

> I strongly suggested he look at gnucap - and will try and
> push him that way again. Working with power-electronics,
> I suspect that  gnucap might actually have some advantage
> for him, especially in dealing with simulations of switching
> converters. 

Hold off.  I am not actively promoting it for now.  I want to 
keep a low profile until some work is done.  I want to attract 
people who can help develop, not others yet, because it in a 
period of rapid change.

Then I want to hit hard.

I think gEDA should take this approach too.

> As an engineer (and student), I am always curious as to how 
> things work, but a long time ago resigned myself to the fact
> I'm never going to understand every detail of how everything
> around me was designed.

Not every detail, but now much of it is completely obscured. 
That's why I like to use simple equipment at the beginning.  It 
gives a good base.

Take as an example .. an adjustable sine wave signal generator.
How many of today's seniors have a clue what it would take to 
make one?  Very few. 

I can do it with one op-amp.  It's covered in a typical 
electronics course, often even with a lab.  It is a real uplift 
to see that the signal generator they have been using has the 
same circuit as the one studied in class.

> What I will say, is that having knowledge of the _principles_
> (and thus  the limitations) of equipment is an absolute MUST,
> especially as things like digital oscilloscopes can mislead
> you if you're not wary. 

The nuances of digital scopes are too much to deal with at the 
beginning. There is a need for simple analog scopes.  So what 
did my colleagues do about it?  Really simple .... defer 
teaching anything about oscilloscopes until the next course.  
So, I had to teach electronics to EE seniors who had
never used an oscilloscope.  Drinking from a fire-hose??

> I hope that you get gEDA + gnucap used in your University. 

No. No support at all.

I don't work there any more.

> I'm fighting the fight here, even if it is just by using them
> myself, and when I can, by working on the code to allow more
> easy integration between the tools.
> 
> Whilst I agree with John's comments that education shouldn't
> rely on wrapping up the inner workings of software, nor do I
> think  that gEDA should dumb down, I do feel that a key
> requirement in  _teaching_ is that the software should not
> get in the way of the underlying lesson.

PSPICE really does get in the way.  It is shocking how much it 
does.  A lot of commercial stuff really does.  That can and 
should be one of our advantages.  I don't mean dumb it down.  I 
mean to do it right.

> Sadly, most students are not comfortable with Unix and
> command lines.

There are lots of tasks that are easy with a command line, but 
impossible ith a GUI.  People who can't use a command line 
can't do those tasks at all.

> If providing a more familiar GUI approach to using the tools
> is possible, and doesn't detract from the core flexibility
> which benefits gEDA, I'm all for it. 

Consistency is important, but that has nothing to do with 
detracting from the core flexibility.  Same goes for error 
handling.  One of the biggest problems with otherwise excellent 
software is how it handles user errors.  It makes sense to step 
back and question the way it is, and make changes.

The traditional approach of wrapper on wrapper may make it look
pretty, but in the long run it just makes a bigger mess.

> It will help us compete with commercial tools - which (IMHO) 
> are not that good in terms of usability anyway.

In terms of ease of use we should be able to be obviously
superior to the high end tools.  In terms of functionality
we should be able to be obviously superior to the low end
tools.  The commercial low end tools are designed to keep you at
the low end, or to sell their high end tools.

I mean true ease of use here, not the GUI hide everything 
approach I see so often.

If you want to see a real bizarre combination of usability, look 
at "Multi-sim".  The graphics are as slick as they come.  The
interaction is outstanding, but some simple tasks are incredibly 
difficult.  It just hides everything important.  I shudder to 
think what would happen if they ever need to do a real 
simulation on a real simulator.

I recall -- I asked students to use a simulator to find all node 
voltages, all branch currents, all collector currents, the 
effective transconductance of all transistors, the AC output 
impedance ...

Looks like a lot of stuff, but it is trivial with gnucap.  One 
of the students insisted on using Multi-sim, and turned in a 
pile of graphs, one for each DC value.  It must have taken many 
hours for what was intended to be about 5 minutes of work.  
With Multi-sim it is extremely difficult to get a simple chart.

Our tools can be the best in terms of usability.  We are not 
far away, but some people think it looks like we are.  Step 
back, hide for a while and make it the best.  Then it will be 
noticed.  The quiet period will help to be noticed, because 
with some care we can make it look like it was that good all 
along.

Maybe we need to make our own line of wiki-books.  I have some 
notes, maybe I can use them to get it started.


_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user