gEDA-dev: Introducing myself

[al davis]

On Saturday 22 March 2008, Adrian Pardini wrote:
> I'd like to work with xgsch2pcb, mainly because I lack the
> skills to write something portable in C and I'm rather short
> on time.
>
> After reading a bit of its code I think I can adapt it to use
> glade, write a plugin interface and maybe some templates.
> When done with the plugin interface I'll start to code
> plugins to interact with simulators and whatever is needed.
> Also calling refdes_renum when necessary might help. I
> believe that xgsch2pcb can be a great aid to help people
> accustomed to other tools in an effortlessly migration.
>
> That's what I think can be done. However, I used pcb and
> gschem for a couple of small projects and I did most of the
> simulations with qucs, so I lack expertise in working with
> ngspice / spice together with the gEDA tools and I am not
> aware of more stuff that can be implemented in xgsch2pcb.

You should check the archives, back less than a month.  There is 
a thread "CSoC" and "GSoC 2008" where Newell Jensen is talking 
about a "project manager".  This is similar to what you are 
proposing, more ambitious.

One reason for discussing things like this in the open is that 
it is better to make proposals that are very different from 
others.  If two are similar, only one will get in.  If they are 
completely different, you stand a better chance if you are #2.

Note the statement at the top of the projects web page: "You can 
use these as fodder for creating your application to Google."   
The suggestions have not been "approved" by anyone.  Rather, 
they are just "fodder".  

Some are too easy for the whole summer, which means you should 
use it as  part of your proposal, but do something else too.  
Some are too much for the summer, which means you need to 
decide on a reasonable subset.

In my opinion, the projects with highest priority are the ones 
that address obvious reasons people don't use gEDA.

With this in mind, I think the translators are the highest 
priority.  The need to hack files to move between our own tools 
is a big turn-off.  An overall project manager is important 
too, but experienced users don't mind (or actually like) 
separate tools.  A GUI won't do any good if the file 
translation isn't complete.  We should be able to use the same 
schematic for simulation and PC board.

This year, I would like to see us concentrate on making the 
system work as a system, while keeping the flexibility of the 
separate tools.