kicad-developers team mailing list archive

[Ouabache Designworks <z3qmtr45@xxxxxxxxx>]

On Mon, Mar 5, 2018 at 9:57 AM, Jon Evans <jon@xxxxxxxxxxxxx> wrote:

> Hi all,
>
> Since my day job involves a lot of engineering planning/timelines/etc,
> I've had this rolling around in my head...
> I started brainstorming some proposed changes to the roadmaps.
>
> I am using Google drive because that's what is easiest for me to play
> with; I'm happy to send patches against the official roadmaps if get some
> buy-in for this.
>
> Feel free to comment (either directly on the doc or by email) with
> thoughts on this.
>
> https://docs.google.com/document/d/1mpxqvxLv497cyfk8KTQijhySSpFxF
> 7ooTtcT_HU86kw/edit#
>
> Basically what I am proposing is to put most of the energy into Eeschema
> for 6.0, with changes to other parts of the software basically being
> "whatever people have time left over for".  Everything else has been bumped
> to 7.0
>
> -Jon
>




My wishlist:

Follow the unix philosophy. All programs must do one thing and do it well.
You solve complex problems by chaining simple tools together. If you don't
then your tool becomes bloated and hard to use and maintain. EEschema is
heading down this path and we need to strip it down to EEschema Lite and
put the ancillary features in their own programs.


Design for BIG data. The rules don't change for big data but they are more
rigorously enforced. PCB's are usually small enough that you can ignore the
rules and still succeed but you will never be able to handle to large
workloads that IC and system designers handle on a daily basis. Why should
you  care? Because you receive deliverables from both of those groups and
if you could offer them a decent open source schematic editor then they
will make their deliverables in kicad formated files.

Your libraries are getting into the realm of big data. Most of the
complaints about libraries on the forum stem from trying to solve big data
problems with small data solutions. Ask yourselves if this would work for
the Library of Congress and if it won't then don't try it.


Create a Pad Mux tool that lets you codesign the IC package and pinout as
part of the IC design team. That is the ultimate in pin swapping capability.




John Eaton