On Friday 03 March 2006 03:10, Peter Brett wrote: > I _may_ be able to get the university to pay me to do nothing > but hack gschem & friends over the summer. If so, what sort > of particular tasks are outstanding and would be accessible > to someone not yet familiar with the codebase?
I think the biggest weakness of the gEDA system is the integration between tools. I would like to see a simulator interface added to gschem, so it appears to integrate with gnucap like Multi-sim or PSpice/orcad. It should be easy to outclass PSpice/orcad. Even a part of this would help a lot. Something like merging gschem and gspice-ui, although I don't think a code merge is the way to do it. The way to do it is to add configurable menus to gschem than can drive the simulator. That is, to issue simulator commands, automatically translate the file format by calling the translator program, etc. Adding back-annotation to gschem, so information from the simulator (voltages, etc) can show on the schematic. Perhaps this can be done by accepting a message from the simulator to update attributes of devices. Open an incoming pipe for an external program to assign attributes, perhaps as lines like "Mout : Vds=2.567" There already is the ability to have lots of arbitrary attributes. If they can be set by an external program, a MOSFET could have attributes like Vds, Id, etc. A net can have a voltage, state, impedance, temperature, etc. Adding forward annotation, so gschem can send such messages to another program, perhaps notifying a simulator that a device was added to the schematic, or a connection was made. Open an outgoing pipe..... If this is too much, a simpler improvement that would be of value is to organize the symbol library. One big problem for beginners is that the symbols needed for any particular design are scattered among many libraries. To see what I mean, find all of the symbols that you would need for a simple transistor amplifier. This is one that could be done by someone who is just getting started, maybe at the level where you just did your first transistor amplifier. The netlister needs to be improved. It probably should be replaced, but to do a good job I think this one is the hardest of all I listed here. It might take the whole summer just to spec out what it should do.
