Re: [ng-spice-devel] Proposal for ngspice development in winter
Here's what is going on with ACS.
I am in the process of moving it to the GNU site, and calling it
"GnuCap", the Gnu circuit analysis package. I have begun to set up
the support stuff on savannah.gnu.org. There are 4 mailing lists,
and CVS will be up soon. You can get the latest at
ftp://ftp.gnu.org/gnucap . There will be a web site on gnu. For
now, you can look at the preliminary one at
http://joule.isu.edu/~aldavis/gnucap/gnucap.html . That part is
almost ready for release.
In development, I have concentrated on the model compiler, which is
now complete enough that the complete mosfet model (all variants that
I have) is now done through ".model" files. The model compiler
automatically generates the C++ files. This was the primary block
for the BJT model. I could have done BJT without the model compiler,
but it is a lot less work with it. The model compiler automatically
generates most of the simulator specific code.
The 4 mailing lists ......
bug-gnucap@gnu.org
Bug reports for gnucap
help-gnucap@gnu.org
Users help list for gnucap.
info-gnucap@gnu.org
Announcements for gnucap. (moderated)
gnucap-devel@gnu.org
Developers list for gnucap.
To subscribe, go to http://savannah.gnu.org/mail/?group_id=562 .
There will be a nicer way to get there when the real web page is up.
For now, you can go to the temporary one on joule.isu.edu, and click
"mailing lists". The lists are archived, but there is nothing in
them yet because they are new.
For me, the summer was mostly occupied with moving and a new job as a
professor. It is a nice location in the mountains. Yellowstone
National Park is close enough for a day trip.
As to C++ ... I remember a few that strongly opposed it, but I they
haven't particapated in a while. Switching to C++ resulted in a huge
improvement in productivity. I am not going back. Also, C++ is one
reason that ACS/GnuCap code size is so much smaller than Spice, even
though the algorithms are more advanced.
al.
Partial thread listing: