> > I am one of the folks who thinks that component symbols should not > > have hidden pin connections. Indeed, this problem is really acute > > when the symbols live in a library, since newbies won't know that the > > pin is connected to GND unless they open up the symbol and look at the > > attributes. > > /me wonders if we should make an exception for the symbols that are > *obviously* denoting hidden connections, like the down-triangle GND, > down-rake "chassis gnd", or up/down arrow voltage rail symbols? > Otherwise you have to explicitly name all those tiny nets connecting > such symbols to each chip.
These symbols explicitly call out their voltage levels, so it makes sense that they be connected to whatever net they have printed on their labels. > The problem of powering multi-gate chips still exists too. Six > vcc/gnd symbols for each hex inverter? We have talked about this at Free Dog. The solution is to have more intelligent slotting of parts. The first instantiation of a slotted part should also carry the power/ground pins. The remaining instantiations of the slotted part would then only carry the signal pins. This requires some mods to the gschem code. Stuart
