Mail Archives: geda-user/2020/10/23/19:34:20
On Fri, 23 Oct 2020, Glenn (glimrick AT epilitimus DOT com) [via
geda-user AT delorie DOT com] wrote:
> 2. Not all spice simulators work the same. It is up to each
> implementation to decide if they want to use the first line, a specific
> .TITLE card, or both/either as ngspice does. For example my reference
> book on PSpice, fairly old so may be different now, requires the first
> line of the netlist to be the title. So spice-directive-1.sym would not
> produce a working netlist for such an implementation.
Hmm. I don't like the idea of simulator idiosyncrasies being coded into
the schematic; this just doesn't feel like the right way to do it.
How about changing the backend logic so the .TITLE directive, if there is
any, is always placed first? Would that work with all simulators?
Are there other cases where directives need to be placed in a particular
order or place?
> 3. Yes it was a minor change, intentionally so, to let me get familiar
> with your process before I offered more substantial changes.
I didn't mean to criticize this; on the contrary, I think it's a good idea.
> 4. re: use of comment. I chose it as it seemed the most appropriate
> without adding a new attribute, and param (my first choice) was used for
> other things. If a new attribute would be preferred I have no problem
> with that. As a gneral plicy use specific attributes would however tend
> to lead to attribute bloat.
If there is no existing attribute that makes sense to use, and you feel
like need to use an attribute, then it's fine to introduce a new one. In
any case, that's better than using a conceptually different attribute like
> 5. I will not be heart broken if you decide to remove spice-title et al. :)
We'll see. I think you have a point, so your contribution is likely to
result in a change in one form or another. And even if it didn't, having
the kind of discussion we're in right now is an integral part of the
process with which you're looking to familiarize yourself.
- Raw text -