Re: gEDA-user: Introduction and some questions/thoughts on gEDA/gaf...

[sdb@xxxxxxxxxx (Stuart Brorson)]

> Hello all!

Greetings!

[ . . . . . ]
> Thus, before jumping to coding, I want to hear some opinions on what I
> will try to do and to learn what people are working on now. If some of
> the things I'd like to see are available (just not from CVS) or almost
>  available and I can help an existing effort rather than start on my
> own.

Excellent!  'Tis always better to augment a working system than
starting afresh, IMO.  

> I am using gschem -> gnetlist -g spice-sdb -> wrspice
> (http://www.wrcad.com/wrspice.html) route for superconductor analog
> circuit simulations. For that I needed to add two symbols, a Josephson
> Junction (useful maybe only to us), and K -- coupling factor (I was
> surprized that it was not there!) + add some lines to gnet-spice-sdb
> to netlist them correctly. I know how I can submit symbols, but should
> I contact Stuart directly to submit a patch or just post it to the
> developers list?

Go ahead and send me the patches.  I appreciate it when folks send
patches (instead of feature requests) because it makes my work easy.
Please supply diffs against the latest gEDA/gaf release, 20050820.  

You can also just send me the symbols and I'll stick them into CVS.
Please let me know which symbol directory they should go in.

> One of the things that were not immediately comfortable to me was the
> fact that it took some trial and error to find an inductor/coil symbol
> which worked with spice-sdb netlister. Would it make more sense to
> keep a library of analog spice-netlistable parts  as a separate
> library (like Cadence analog lib which did not come with, e.g.,
> symbols for batteries, but for voltage sources which sure as hell
> could be netlisted to spice, spectre and hspice)?

There is a "spice" directory, which holds SPICE specific components
(VCCS, Voltage, etc.).  There is also an "analog" directory, whose
components (standard linear passives, IIRC) should all netlist
correctly for spice-sdb.  Which inductors failed?

I agree that the symbol directories are a little disorganized.
Perhaps creating a separate "spice-component" directory holding
garuanteed resistors, inductors, caps, transistors, diodes, etc. might
be a good idea?

> Is there a way to get access to guile interpreter from within the main
> gschem window? (I know that I can run gnetlist with -i flag). Besides
> the fact that having a scheme interpreter in a separate window is
> convenient even to run as a calculator, hopefully more scripts will be
> written if interactive session becomes available. Those who've used
> DFII/Skill will know what I mean... ;-)

I dunno the answer to this one.

> Also, how hard would it be to add buttons to the toolbar from guile,
> with attached callbacks to run guile scripts (e.g., some (or (system
> "gnetlist") (system "spice")) , you know...).

We'll never know until you submit the patches implementing this!   ;-)

> Hierarchical netlisting support... I understand that gschem is mostly
> used for board-level design (where there are only two levels of
> hierarchy: package and board); Stuart's example of a "transistor with
> parasitics" is on the other end of the spectrum, but still spans only
> two levels. For anything more hierarchical that this, two things are
> essential, IMHO: easy automatic generation of symbols starting with
> pins on the schematic view and allowing gnetlist recursively netlist
> subcircuits encountered in the top-level schematic. I'm wondering if
> anybody is working on it, I'm eager to help.

If you create a lower level schematic called "foo.sch", and an upper
level schematic called "bar.sch", then you should be able to do this:

1.  Create a symbol called "foo.sym".
2.  Place foo.sym in your top level schematic bar.sch.
3.  Attach the attribute "source=./foo.sch" to the symbol foo.

Then, when you netlist, gnetlist is supposed to do the right thing.  I
don't recall testing it, however.  Here's
Ales' post answering exactly this question:

http://www.geda.seul.org/mailinglist/geda-user7/msg00005.html

BTW:  One known problem with gEDA/gaf is that busses aren't passed
through the hierarchy.  If you were looking for a project, fixing the
way busses work in gschem and gnetlist would be a great service to
everybody! 

> Finally, as good as user interface is (love the keyboard shortcuts!),
> is there a way to map keyboard arrows to panning actions? ;-)

You can remap your entire keyboard using the gschemrc file system.
I believe the required incantation has been documented in teh docs
distributed with gEDA, but I could be wrong.   In any event, you can
study the system-gschemrc and system-commonrc or system-gafrc files to
see what is necessary.

HTH,

Stuart